Figure 1. Woodland surrounding fields (a), its expansions (b,c).
Held on Zoom: Friday 21 October 2022
Themes of Discussion:
I. Overview of discussions and What We Learned.
II. Proposal for a Dooyeweerdian Approach to AI
[In this, the 6th RLDG discussion of discussion on Artificial Intelligence, JC led us in six major questions aimed at compiling an overview perspective. This then led into a proposed two-stream approach, covering the 'social' and 'technical' sides of AI. The purpose of these discussions is to contribute to developing a Christian and Reformational perspective on AI.
Summaries have been added into each section. ]
Present: JC, SJ, NB, AB.
Apologies: TB
[This discussion was hosted by, who typed notes as it progressed, recorded it and then converted the notes into an annotated transcription (6 December 2022). AB adopted two roles in doing this, (a) of editor ("[Ed. ]"), e.g. giving links to other material, adding headings, adding "***" to important points, adding notes in square brackets that explain things, and attaching unique labels to statements for future reference (currently only marked as "[]" but not yet assigned unique labels); (b) of contributor to the discussion ("[AB: ]"), inserting responses subsequent to the discussion, or maybe explain what he had intended to say but left unsaid. The purpose of these is to take discussion further, and not to dominate, and include some where he might even criticise himself for what was said on the day!
In the transcript I have tried to put down the exact words spoken but, when there are falterings, especially which repeats like such as "a little bit ago, a few weeks ago", which add no meaning, I omit some words without indication. However, where I judge they might indicate something meaningful, such as uncertainty in the speaker that it is important to record, I might leave them in.
]
[recording started]
# JC: So, AB and I chatted a little bit ago, and we've been just discussing. And some of us have looked through, read through, listened to the recent Reith Lecture on AI and its implications.
# The message came a few weeks ago about "AI - Is it?" []
# What does it look like when we worship a new technology. []
# JC: So, what I'd like to do is conduct this. I used to teach adult Sunday School, and I'd like to conduct this similar to that, where it's more open discussion, reflection. So I can really understand each of your own, and my own, and our perspective on certain things. And then that will be the first part of the conversation. Just so that we can have a clear understand of where each of us see things. Broad questions. []
# And then the second part will be more of - I will say prayerful reflection and just kindof listening to everyone. And listening to the industry at large, different valley?? folks, different investors that I get to speak with, and just kindof / a hypothesis or recommendation for us in our daily lives so that can move through towards God's principles of benefit for all. []
# So that's kindof what I'm proposing.
# So the first thing I want to do is share this, if it will open.
# I will go ahead and
# chat: 00:11:01 JC: https://docs.google.com/document/d/1Be9IbWPQgh1OyGpa4N0Kskz3eYIZ3ii-a8KnwD6_wd4/edit
# AB: You should be able to share screen if you want. # JC: OK, let me know it it opens in your computer so you can see it.
# [Ed. JC tried to shared a doc on screen]
# NB: It wants me to request access. My gmail is [Ed. redacted]. I chose it fifteen years ago when I didn't have a lot of time and that is the word I came up with!
# JC: OK, everybody go ahead and request access. [some time in getting access for everyone; quite comical to listen to.]
# SJ: I'll send a request. JC: Request granted. Now I'll move on and if at any point it's not clear what I'm asking, you have the words as well. [AB mumble to access the document].
[Ed. document shown on screen, which JC would then write into]
# JC: If at any point it's not clear what I'm asking, you have the words as well.
[04.10]
Contents
# JC: What have we learned regarding AI and the use of this new way of process information? What have you learned as a result of these lectures? [AB found the questions]
# JC: Who wants to go first? SJ?
# What have we learned regarding AI and the use of this new way of process information? As a result of these conversations or this lecture?
[04.50]
Contents
# SJ: One thing that immediately comes to mind is that AI is able to process millions of rows of data faster than human. []
# JC: Thank you; anything else?
Contents
# SJ: I think it was in one of the last conversations we had on that, is that the kind of knock on effect on humans was the lack of / I'm thinking it is direct effect, but it's more of / Human becomes more passive when working with AI, in terms of their critical thinking and curiosity. And I'm talking about lay people; I'm not talking about the data scientist; lay people who are given AI-based applications to do the job for them. [] ***
# JC: Thank you very much.
[06.35]
Contents
# JC: NB?
# NB: I'm surprised, in all of these discussions, both in the Reith Lectures and in what we've had, how it makes so clear that we have different conceptions of what intelligence means. That, we all use that word but we don't use the same thing by it. []
# And even the notion of what it means to be human has always been slippery. []
# But AI discussions bring that to the fore in a way that is very visceral, that we don't really know what we mean when we use these words. [laughter] We don't mean anything very specific / []
# JC: Are you saying that our actual, in the lingual aspect, refers, has the retroposity across, different signals. It is not weighted the same across different aspects, on the dependent aspects? Correct? []
# NB: Different people will then concentrate on some of the different asepects, so to speak. []
# NB: I noticed that in our discussion of the drawing program, the Wombo program, that there is in some sense there is disagreement about whether the computer is / The question of "Is the computer creating art?" - we cannot even discuss the question because we don't even know what we mean by the terms. Or at least we don't agree on what we mean by the terms. []
# JC: We don't have rhetorical agreement on actually the meaning of terms. We have a central post-modernism critique. # NB: Yeah, that's right. []
# NB: This is not a new insight but I think AI brings it to the fore in a way that is rich. # JC: Yeah. # NB: It forces the discussions. []
[08.25]
# JC: I love that; thank you. Thankful for the philosophical education of ??Grove City??.
# And mainly Levi-Strauss (are you familiar with Levi-Strauss? # NB: Only vaguely.) # JC: So there's an / AB and SJ, you might be / the language-mind-reality philosophers, Derrida, Strauss, and some other folks, really kinda hit into this new form of (like) "Our meaning must be tethered to something if there's meaning at all." But we cannot agree on some sort of tether; we cannot agree about it. []
# So, what Dooyeweerd provides, is kindof multi-aspectual tethering points, that you are bringing up NB, that we have such unclarity on what this means, on these other aspectual points. []
# I think it's ultimately because our pistic aspect is not aligned with so many other people. Yknow, we don't have the same worship trajectory. We don't have the same worship point. And when we don't have the same worship point, it kinda skews where all the aspectual weight is. []
# JC: Is that clear statement to you guys? # AB: Yeah. # NB: I think so.
# AB: That's nice, actually, "worship trajectory" is nice.
# JC: NB, do you have any other insights that you would like to share on this question? # NB: Not at this moment.
[09.50]
Contents
# JC: AB, what about you?
# AB: "What have we learned regarding AI and the use of this new way of process information?"
# Because I have transcribed the discussions, I've gone through and I've learnt a lot. Or maybe some things were not so much learnt but, what I knew before has been modified. []
# AB: One of the interesting things is:
# I used to work in Knowledge Representation AI. [Ed. See 2nd RLDG Discussion indented paragraphs za281 on KRAI] Of course, most people don't know what that is, because it was in the 1980s, and then it 'died' and machine learning replaced it. []
# But I'm seeing more and more the similarities and differences between this [knowledge representation AI, KRAI] and machine learning [MLAI]. []
# JC: Can you expound on that, where some of the similarities and differences? # AB: In some way I've done so in the transcriptions. When I write something down, it disappears from my brain. []
# [Similarity 1] So for example, you can see both of them as: you need to capture some knowledge and put that into a machine in a way that the machine can operate according to that knowledge. That is the kindof fundamental thing about AI. [] ***
[11.50]
# [Difference 1] But the way the two technologies do it is different. []
# KRAI: We interview human experts [to elicit expert generic knowledge that is fairly generally applicable]and we then represent that knowledge [elicited from experts] in some knowledge representation language. []
# In MLAI we don't try to capture general knowledge as such. We just give it a lot of data and let it work out the patterns in the data. []
# So, I would not say "GP AI" I would say "GP ML". # JC: OK, thank you. In fact I won't say it's actually GP, General Purpose AI. # JC: just ML? # AB: Yeah. []
[Ed. "GP" = General Purpose, i.e. with knowledge not confined to one domain; I don't know why AB mentions GP there, because on the recording at least it does not seem to have been mentioned before. Maybe it's a reference to an early discussion of it, or maybe zoom just failed to record it clearly; there was a few seconds earlier when JC and AB were speaking together and their voices got mixed up.].
# AB: [Similarity 2] And in both there are limittions. []
# [Difference 2] The limitations in KRAI is that / is basically tacit knowledge, the knowledge that you [or experts] know but cannot express, or don't express. And so, unless you have got a good way of obtaining and expressing the tacit knowledge, which I think Mike Winfield did give us, using aspects, you will miss a lot of the knowledge that's important, [] ***
# because it is the knowledge of exceptions. If you ask someone for their knowledge, they will give you the straightforward knowledge, the things that are generally so. and won't give you the exceptions. So you need a way of obtaining the exceptions.
# Basically, the exceptions are hidden in the mind. []
# There are various ways in which things are tacit. I worked with a PhD student [Alex Kimani] on different aspects of tacit knowledge - I'm trying to get him to write a paper. []
# [Difference 2, continued] With ML AI, that is not so much of a problem, because / []
# NB: Does it mean that your dataset has access to those exceptions implicitly. # AB: yes, that's right. []
# AB: I'm assuming you've got a dataset / Because the dataset, if you like, shows what actually happens in reality - or, has happened in reality - which will include, and which will operate according to, the tacit knowledge (or what was, the tacit knowledge) as well as the explicit knowledge, together. []
[AB: Or, better: according to the exceptions and the straightforward knowledge, together]. []
[15.40]
# AB: But there are two limitations in that. Because, as NB says, that is a big / there is an assumption there. One limitation is / there's three. []
# AB: [Similarity 2] One is, is your dataset big enough to be able to capture all these rare exceptions. That's a technical one [reason]. [AB: Similarity to KRAI in that you need a domain expert with sufficient expertise.] []
# [Similarity 3] Even it is, there is the issue of selection of variables that are meaningful. And often, the people who select these variables - this is a claim that is not substantiated although I don't know whether it's been studied - the people that select variables [for training the ML system] do it according to what is meaningful to them, and they might omit whole aspects that are actually relevant but they just don't bother / they don't realise they need to have those variables in the data. Which is, again, almost a version of tacit knowledge, a different type. []
[AB: In fact, it could make the problem of missing knowledge worse, because, with KRAI there is at least the conversation between expert and knowledge engineer within which the lack of whole aspects might become noticed and dealt with, whereas in MLAI there is no such conversation. ] [][AB: Similarity with KRAI: Both KRAI and MLAI depend on the analyst selecting material that is meaningful from what is not, into separate concepts.] [] []
[AB: These are bias of the developer, referred to later] []
# AB: [Difference 3] And a third limitation is that MLAI always looks back to the past, never to the future, because it is derived from past / the data is always data from the past. Whereas KRAI, if you've got a wise domain expert and good knowledge engineer (a good knowledge engineer can say "Yeah, but what if?"). []
# [Difference 4] Actually, there's a fourth limitation a well. The fourth limitation of MLAI is that you cannot yet (I don't know that you ever will be able to) you cannot yet see what the knowledge is that is in there; you cannot check it; you cannot get it to explain itself. []
# JC: Is that because tacit knowledge formation comes within the group analysis? Like, each dataset has a new form of kindof knowledge aperture, so you don't set the rules, the laws, within that? []
# AB: I don't think I mean that. It's that the knowledge in the machine is held as the weights of links between neurones or neuronic cells. []
# NB: It's similar in a way to the fact that even if you dissected my brain at the neuronal level, you wouldn't be able to tell what I was thinking, or why I came to some decisions. AI ML is similar, that the way that the knowledge emerges from the weightings is opaque. # AB: yeah, opaque; that's a good word for the knowledge. []
[20.00]
[AB: This may be explained in Dooyeweerdian aspectual terms. KRAI operates at the analytical aspect, with distinct concepts, distinghished by the knowledge engineer. MLAI operates at the psychical aspect, similar to pattern recognition that occurs in animals. Dooyeweerd recognised that animals can distinguish their mates from others, etc. but do so psychically rather than analytically. Humans make analytical distinctions. It is because of being composed of distinct meaningful concepts, why KRAI is transparent while MLAI is opaque. ] []
# JC: Thank you, everyone. Can I move onto the next one, because I have nothing to add.
Contents
# JC: Like, has struck you? Like, SJ, you have mentioned a little about the implications of General Purpose AI with leave?? the general population kindof like that / what word you used, SJ, lack of critical thinking. Anything else strike you, about these new technologies? Anything? []
# SJ: One characteristic of them is that are quite exponential in how they grow in terms of the application which is developed on them. That's one thing. []
# SJ: Second is, if you just come and look at them as the one who / someone who is amazed by how wonderful AI and its applications in terms of solving problems, it is very easy to fall into what we call idolatry, worshipping it, which is a quite dominated view in every government conference and academic conferences. Looking at what it does; "Look how amazing it is." [] ***
# JC: Well said! # AB: Absolutely. [laughter]
# SJ: Because we get overload, a lot of news that "AI does this, AI does that. AI does X." There is no chance to really look at the downfalls of them. []
# I know government / This is interesting: I have been to the digital roundtable with NorthEast Chamber of Commerce in the north-east of England. I really mentioned / because everyone was talking about "Lets share the success stories of AI, what AI does for the business." []
# SJ: And I said, "Wouldn't it be beneficial for our audience, like the SMEs in this region, to know about the failure stories as well?" # AB: Well done! It was not welcome. [] ***
[23.30]
# JC: [laughter] It wasn't welcome! # SJ: Because the focus is hush-hush about failures so that / The funding bodies, the specialist funding organisations, they want to hush-hush about failures. []
# JC: It mirrors many technological adopting curves, in many new spaces of any technology?? across the board. "We have to keep it quiet so the market can adopt it and continue to pour cash into it." Pretty easy to see it there. But that is pretty well-said. Thank you. []
# NB: It is the nature of idolatry. Idolatry always takes something good from Creation and elevates it to an inappropriate position. [] ***
# So the challenge of not putting AI on this pedestal, not trusting it in a way that our trust ought not to be there, while at the same time enjoying the fruits of God's good Creation, which includes AI, I don't know how to do that. []
# In a world full of idolatry, to point out any of the good benefits of the idols is contributing any more to the idolatry, and I don't want to do that. At the same time, it doesn't mean that the gold and the wood that the idols are made out of are in themselves evil. []
[25.00]
[Ed. See also More on Idolatry.]
Contents
# AB: That's interesting. You say you don't know how to do that. I think you are saying, you don't know how to, if you like, to prevent the idolatry. I'm not sure that we are called to prevent idolatry. []
# Um. Now, this is a theological thing. I was struck as a few years ago i read the beginning o Amos. Amos, chapters 1 and the first half of chapter 2. And there the Lord God slams seven nations, and says "Woe to ..." and so on, and "this is going to come upon you." And the last two is Israel and Judah. And the others were nations around, like Edom and Moab and / And what struck me was, what God condemned them for was different. God condemned Israel and Judah for not obeying His law and for their idolatry. But He did not condemn the other five nations for their idolatry; He condemned them for what we might call war crives. []
# Now, that made me think, "If God / If idolatry is important to God, why did God not condemn them for idolatry, because they would have been worse idolaters than Israel and Judah. And the answer that I came up with, or came to me, was that Israel and Judah represent God and therefore they should be seen as non-idolatrous. Whereas the other nations, if you like, could not help their idolatry; the Law had not been given to them. Israel and Judah were to be a beacon, a light to the world that they [other nations] could turn to if they wanted to get rid of their idolatry. []
# And from that / that was a signal moment. It changed my view on the importance of idolatry - that it's not something that we are / unless the Lord specifically asks us to, we are not necessarily to fight idolatry. []
# Now, there are situatons where we must, bot, in general terms, what we are to do is to offer Christ, and offer a beautiful Creational view on things. []
# NB: Is there a distinction you would make between fighting idolatry in our culture and fighting it in our church? [] ***
# JC: Ahh!
# AB: Yes. I don't really want to get into that because we are on an AI thing. But fighting it in the church is like condemning Israel and Judah for it, because the church represents God, Christ. []
# JC: That's an important point. Let us examine that later on. Thank you for that note, AB, what you just said. []
# JC: But what you also just said prior to what you just said is what I think is the unique beauty of this gathering of people, digitally. Being able to listen to you and learn from you and others, to see how we navigate our proper place as Christians in the world today, in our specific spheres, and acknowledge where we should and should not, and accept those restrictions and directions from each other. Thank you so much, AB, thank you. # AB: Right, OK. [] ***
[29.25]
# AB: That's an aside; let's get back on to AI!
# JC: Why do you think we should care /
# AB: I was saying, what was NB saying? He was saying "I don't know how to do that. But the gold and wood the idols are made of is not in themselves evil." So that's the point we need to look at; Great. []
# JC: So it's like not looking at, like, the newest Natural Language processing algorithms as the most important thing that man's ever had. Like, to know that they are simply tools. Yknow and we should not dismiss them as effective tools. [Instead] we should know humbly and meekly, effectively, with appropriate love. []
[30.05]
Contents
# NB: In the sense of, "That's so cool!" - I'm a computer nerd, a computer scientist. When computers do new things we previously didn't think they were capable of, I think we are allowed to enjoy that, sortof revelling in "Hey, this is nifty, this neat, this is cool!" # AB: Amen! [] ***
# NB: And not taking the next step of "And therefore we should do this everywhere, all the time." []
# SJ: It leaves me thinking, that even though I like what AB said, we are not called to fight idolatry. Or, you know, the new innovation, finding as computer scientist, we can enjoy. []
# And, I think it is a matter of attitude. If we are enjoying something a lot. If the attitude is one of bringing peace on earth, you are allowed to enjoy your innovation on computer. You can enjoy it, because it is driven by compassion and a righteous attitude. [] ***
# It might be seen by idolatry from outside by others, but if your heart is in a/the right place, the outcome is not necessarily really bad. So that's why we need to be cautious of not fighting with idolatry, as everyone who is enjoying the AI so much are / yknow they are doing out of selfishness or self-promotion. Yknow, like the keynote speaker of x y z. Dyou know what I mean? To be mindful. []
[AB: Interesting and important discussion. With the distinction between ethical and pistic aspects, we may perhaps see two dysfunctional attitudes here, often intertwined. Selfishness and self-promotion, ethical aspect: using AI to further my own career, kudos, etc. Idolatry, pistic aspect: overweening commitment to AI, ignoring other aspects and especially refusing to allow discussion of its disbenefits! ] []
# JC: Well said.
[32.15]
# JC: On the next question I do want to say ...
Contents
# JC: I do want to say, kindof as a point of narrative input: A lot of that has already been shared. But if there anything specific you want to share, like Why do you think we should care about the use of AI within human affairs? If there is anything specific, like "This needs to be said", please do share, anyone. []
[Ed. See also Idolatry. ]
# AB: If I may, I'll come in. I was about to add to the idolatry thing. Having said, that I don't think we are called generally to fight idolatry, I think we are called generally to, if you like, to warn humanity against the harm of idolatry. That, if we idolise something, then things will go bad, as we have seen in economics. And so, that, if you like, is qualifying, modifying what I've just said. [] ***
# AB: It's not just fighting idolatry in the church, because we are God's representatives, it's warning, lovingly warning against idolatry, in lots of different ways, in the world. So the Moabites / Israel and Judah should have gone to the Moabites and so on and said, "Look, because you are worshipping Molech (or whoever it is), your society is being wrecked; why don't you come and worship Yahweh with us?" That kind of thing. []
# NB: Parallels in AI there are somewhat obvious: that we can see the dysfunction that will result from people unquestioningly implementing AI "because that's what everyone is doing." []
[34.30]
[AB: Summarise that. Three points about idolatry and our responsibility as Christians with regard to it:
- We are not called primarily or generally to fight against idolatry in the world.
- We are called to fight against idolatry among God's people, because they / we represent God - though we do so with love and with rather than for God.
- We are called to lovingly warn the rest of humanity about the real dangers of idolatry for their society and culture, and individuals.
The latter involves taking into account not just the direct, visible impacts of idolatry, but the hidden ones by which it has retrocipatory impact on all the rest of our living, thinking, doing and saying. ] [] ***
[See also Deifying AI.]
Contents
# SJ: If I may add, on the use of AI. I think we need to also care about what do we mean by "use of AI". Again, "use," like "intelligence" or "human", it means different things to different people. []
[Ed. For some reason, nobody picked this up. ]
[Ed: Maybe SJ was meaning there are several kinds or ways of using AI. So I emailed him and he and NB replied with the following.SJ: "I guess below is what I meant:
Use as direct involvement with the algorithm: An example would be a data scientist wrestling with multiple competing algorithms to find an optimum model.
Use as sem-direct involvement with an algorithm: An example would a robotic nurse in a hospital that learns through human requests and interaction through an interface.
Use as non-direct involvement with an algorithm: An example would be when we sit in the bank clerk's office, and she checks our credit through an AI-enabled application; she lets us know the outcome, and then we decide based on how she interpreted the insight given to her by the algorithm."
NB: "What I think you're describing is a sense of "proximity" to the actual nitty-gritty of the algorithm. In the first case, you have someone working directly with the algorithm itself, whereas in the last case, we have someone who's treating the algorithm as an inscrutable "black box" and just accepts the answers as they come out. ... people accepting the results without any understanding of either the algorithm or the training data."
SJ: "... not thought of sense of proximity before. I think it is a better term than direct or indirect involvement." Two people in a firm say, "We now use AI to run preventive maintenance", but understand it differently: The CEO of an oil and gas company and a data scientist of the same company. ] [] ***
[35.00]
# JC: Gotcha. Thank you.
# JC: Anything else?
Contents
# JC: This is where we'll take a rhetorical turn. Who isn't familiar with Latiin phrase "Quis custodiet ipsos custodes?"? # AB and SJ at least were not. # JC: It is, "Who watches the watchmen?" Seneca, right NB? I just know the phrase. It's such a kindof oonstant tension on / There are laws, but who make the laws, but who watches over those who make the laws? []
# It is an undertanding that power without any sort of checks, balances, or anything, obviously leads to exploitation and worship of that sense of power. []
# It is a continual kindof tension in society, and everything. Who watches the watchers? Who guards the guardsmen? Then you can apply it, like / It takes a step to be, like, to "How could we do that?" And that's kindof where I want the mind / SJ do you undertand that statement now? # SJ: Yes. # JC: I did not want to just breeze over that. Thank you for speaking up. []
# SJ: Is it Greek words, or what language is it? # JC: It's Latin. I think Seneca; I'm sure a quick google can bring that up. I think it's the only Latin phrase I can write out. []
[36.50]
Contents
# And that is the touching off point that I want to bring out as just the discussion point now. []
# JC: What is the new way we can encapsulate AI, so that we may properly steward the spheres in groups that we are within, in our daily lives, for the glory of God and building the Kingdom [of Heaven]? []
# This is what I've kindof been reflecting on and kindof thinking through, after listening to stuff. []
# So I'll just share a couple of statements, OK , and then I would like to hear your reflections. But I want to be respectful of NB's time to leave. I do want to hear NB's feedbackl on this before he skidaddles. []
# That's the only reason I'm being more rapid at this stage: I want to respect / [soom multiple voices].
Contents
[See also Idolatry.]
# JC: What happens if we deify technology? So, this is some of the things that we've been talking about. We're deifying, we are putting AI in the place of worship. []
# What if we simply remain to see God in the proper role God is, and simply see AI as a tool? [] ***
# Kindof like what you are talking about, NB: What if we just see it as a tool?
# Tools can break, see??your your point, SJ. Where they don't work, we should talk about them; where they do work, we should go, like, "Oh, that's cool!" but not automatically apply it to every situation. We should be wise as to how we apply it. []
Contents
# So this is where I will say, "I suggest AI should simply be understood as a shovel in time." AI is a shovel in time. Something that digs large processes of formerly cumbersome calculations away, in rapid mesmerizing speed. So we are, like, "Oh! That's awesome. But it simply a shovel in time. []
# So, lowering the time threshhold on many potential applications. []
# But when you've got a ??decrease??deep-freeze?? for ??tradition took?? a lot of time. []
[39.00]
# So, if we think of it that way, and think of the process of that digging, I have one more analogy to introduce. []
# JC: In a way, it's a new form of ??injden?? [engine??]. And this is getting towards where I want to apply our thought. []
# What if we see AI, as a system of digital processes that work together to bring about a new, vastly faster way to travel from unclear data to real-time, appropriately-determined and applicable information for use within a given context. [] ***
# Kind of an internal combustion engine of data. So it's just a new tool, right?
# So then, much like the steam engine transformed salt and other forms of mining, the truly revolutionary use was found when it found itself with four wheels and travelled on existing roads. And those roads were designed for horses, carriages, they were not even designed to withstand the new weight and pounding and the way in which the engine worked. []
# So, due to the development of the car and popularity within certain segments of society, new forms of road were developed to enable the use of cars, and more and more playthings for the rich. []
# So these new forms of roads are what I think we should be focused on as a group, to usher in more discernible faster ways to truth and flourishing, shalom in God's Creation. It's the roads, it's the pathways, it's the proper application. []
# It's not worshipping, as in everything, it's the proper application, like, you cannot put AI in biological trials that need human forms. Because of absolute / yknow, you cannot / AI in a human body. So you shouldn't apply, like, assuming that the tool should work and it will all solve all the human problems. []
# But before I talk any further, NB, do you have any things to say to that?
[40.55]
Contents
# NB: In your analogy, not just concentrate on building the roads but also on the traffic laws, and the societal expectations of how vehicles work. [] ***
# Yknow, the idea of the automative revolution is a powerful metaphor, because it did change everything about our lives. And looking back at the 1910s, 1920s, could we have done the introduction of the automobile in a way that would be better for society? Yknow, fewer / less surburban sprawl for example. Yeah. []
# And so, looking ahead to the similar question of AI, as we build an infrastructure that allow us to keep AI in its lane, in a sense, it's Not only the physical infrastructure that enables, but also the societal infrastructure that constrains. [] ***
# JC: Mmm. That's brilliant. And it's like what should be. And using those terms appropriately, I think in the church we shouldn't move away / we shoudln't. We should appropriately understand / we should / we are recommenting (AB's point), "This is the problem of idolatry, this is what happens when you bring in idolatry." So we ought to share that information, so what should be the proper societal constraints? []
# NB: As an example - and I do have to wrap this up - I've think we are probably not too far away from having a zoom call with an artificial agent where we cannot tell that we are talking to an artificial agent instead of a real human being. [] ***
# And I'm troubled by that, in that I don't want to waste my compassion on something that is not human. [laughter] So I do not want to not know whether the entity I am talking to is a real human or a simulated one. []
# And that's yknow a societal level decision. We could decide as a society what to / what the expectations are when you are chatting with an artificial agent. Do you make them look intentionally robotic, so that it's clear what you are talking to, or allow companies to try to trick you? []
# JC: NB, that's exactly / that example. But these are the things that I think our groups and our influence, and where we are, can care for those around us. [] ***
# JC: Thank you. Thank you, NB, and I know you've got to scurry.
# NB: So I look forward to reading the transcript and anything else you guys come up with. And I will see you around next time. Have a great day, everyone.
# JC: God bless.
[NB left]
[43.50]
# JC: SJ, do you have any reflections on either what NB said, or what there?? or do you want me to repeat anything else, or anything?
# SJ: So, we are looking at the last question, correct?
Contents
# JC: Like, how can we best use our lives, our care, our love, our knowledge, our points of contact, our life, to best help? []
# SJ: If we are able to detect in our daily life those moments we are / (I mean, I'm not sure how it is relevant to this question) but what I am trying to say is, of course I agree that we should avoid AI to be at the place of worship. []
# But for that to happen, we need to value the precious opportunity we have to in this life on earth. That we have / we are able, as human, to do certain things, and we should let AI to serve that rather than we serve AI. [] ***
# So, we are able to have be compassionate toward others, love one another. Yeah? We always need to think how AI can perhaps enable that to be expanded. [] ***
# Rather than AI override that with some amazing things it does. []
# So we need to really [be] thankful for being human, for how we are created in the image of God. And then perhaps that's the better way of thinking it. []
# If you are not mindful of that, that can give its place to the wonderfulness of AI [i.e. worship]. []
# So in our daily lives, I think we should remind ourselves time to time how wonderfully we are created. [] ***
# AB: Yeah, I like that. It reminds me of what Paul says in Romans 1, where he says the beginning of problems was because people ceased to thank God and treat Him as God. That's what I hear in a beautiful way in what you are saying. []
[47.40]
# AB: I notice the writing has stopped, Or at least on my screen. [To jC, who had been writing into the shared document on screen]. # JC: Well no, I'm keeping notes on the second part that I shared on. I have all the notes and I'm writing everything still. []
Contents
# JC: Let me share one more thought. statement. Then, AB, I want to hear your thoughts. But then I just wanna get this out, so that it's a part of our shared discussion as well. []
# JC: Our work, which can humbly embrace with meekness / and that's meekness in the terms of kind of what you're saying, SJ: meekness is knowing how to use a sword but keeping it sheathed. Not having to throw it out all the time and show how often we are "Look, I can slice up an apple." []
# It's like, "Cool, you've got a sword; keep it sheathed. ???ill Be a human, recognise how wonderfully you are made: that compassion. I love your statement, "How do you expand the compassion" to other humans through AI. I love that, it's beautiful. []
# What we could use is our rhetorical pathways to influence the adoption and use, with shalomic parameters, this engine for the grace of all. []
# That's how I say what you've said. How do we propel the proper use of this tool for the grace and compassion towards all? [] ***
# AB: Can I ask for clarification? You said, "use our rhetorical pathways"; what do you mean?
# JC: So, writing, speaking, influence. Like, actually taking an active role. Kinda like how SJ is, in this Northeast Chamber of Commerce, at least trying to speak lovingly, saying "Yo, we should at least talk about the problems." And they may shut you down, and you love them and not push against them, and be a jerk. You at least speak. And say, "Hey, think about this." [] ***
# And the way that we can continue to refine one another is: [discuss] the better way to say, "Hey, you should think about this." []
# So that's what I mean by being very conscious and mindful. Yknow, in the sphere that we operate in, understand where we are embedded, understand where these are appropriate. And understand where we want to state, "Hey, what about this?" []
# JC: Because I fully believe in this as a shovel or engine. It just drastically decreases the time. That bunch of time does open up a new level of human flourishing, exploitation, mainly just on the science research ??? the biological science. Yknow, like, human trial sciences. []
# It's just enormous! And it has applications for every other field too. []
# But it's the wisdom with how we should prioritize and what our meta-ethic is. []
# And I think, SJ, your statement of how we can share compassion to people, first. Could be a great, like, starting path. []
[51.00]
Contents
# JC: Now, we can walk in faith, not in worship to a tool, yknow, and just /
# New formulation of data processing can provide goodness. That's what we / how it can provide goodness. [] ***
# SJ: I can link that to what AB said in terms of four limitations on machine learning. []
# If you go back up, one of them is that ML project is subject to the bias of the developer. [] ***
# Yes, so it is the tacit knowledge of the data scientist that could choose and select which variable to take, which variable not to take, which one to more emphasise; what is meaningful to the data scientist, actually. So his or her tacit knowledge, the frame of mind is affecting the process. []
# SJ: [Example] So imagine that frame of mind, its centre has compassion for the rest of Creation, for human being. That would affect the type of selection he does or she does, of the variables. Rather than someone who is thinking, "This is a 6-month placement for me. This company let me just do some project." []
# It's totally different frame of mind! [] ***
[52.50]
# AB: I like that.
# JC: I heard a line the other day. Although I'm not being careful to not say "save" or "rescue" humanity, I like your word "compassion", to show grace. Because pretty much, every war in history has been started by one party thinking they are going to save something else. So they're, like, if they're going to kill you you. And so none of that. It's not about rescuing or saving; it's about how to show compassion, how to show grace. []
Contents
# JC: Maybe we should take, in our lives, pray for all the data scientists in the world, for five minutes, and be, like, "Let them have compassion in their minds." Like, if we think we are Christians, and we think prayer works, maybe that's something that should be part of our daily prayer life. [] ***
# I mean that, honestly. # SJ: Absolutely. I am thinking, if we believe every human being is created in image of God, and they have the potential of being like Jesus, in them, that includes data scientists. And it is so wonderful, if we just for five minutes rejoice in their potential. That type of potential. []
# SJ: And see the world that these data scientists tap into that potential to develop a new mindset, a new world[view?], even by compassion and love of Christ. That would be a different outcome in the world. []
# JC: Ahh, that's beautiful!
[54.50]
# JC: AB, what are your thoughts?
# AB: I'm mainly listening. I'm loving what I'm hearing from the two of you, and from NB. []
# AB: Bur let's see. "What is the new way we can encapsulate AI, so that we may properly steward the spheres in groups that we are within, in our daily lives, for the glory of God and building the Kingdom?" What is a new way we can encapsulate AI? Well, a lot has been said. Let me see if there is anything more that has not been said. []
# AB: This idea of compassion and / rathrr than saving the world, having compassion. I like that. I tend to be a person that likes to save the world. And especially in my green activity. And sometimes it can be a little too much on people, rightly so. So, compassion and its ??? []
[56.30]
Contents
# AB: However, in terms of AI, one thing I - I'm sitll thinking this through, but one thing that's important to me is when I am representing reality of the world in computer terms [AB: whether this is KR AI or ML AI]. I don't do much of this now, because I am sortof at the higher level of use and economics and these sort of things. But / []
# There's two levels at which I can be representing the world and how the world works in computer algorithms. []
# So, for example, if I am designing an AI for choosing routes in a Satnav (just to take a simple example), that is an application, a particular [type of] application. However, it requires good knowledge of the spatial aspect of reality, in order to plan routes. [AB: AB should have said the kinematic aspect.] []
# And probably some knowledge of what is meaningful to humans in choosing between routes. And my car, for example, has four options. One is fastest, shortest, most ecological, and something else. But what about the most scenic, most beautiful? What about, "I want it to ecological under certain conditions, for example I only go 50 mph on a motorway, instead of 70 mph which are ??consumed??. And all these sort of things. / Got mixed up there / []
# The laws of the spatial aspect are the fundamental things. What humans find meaningful is the application things. So those are the two things. []
[AB: I would like to modify that now. First, it should be the laws of the kinematic aspect rather than the spatial but, since the kinematic depends on the spatial, we would need spatial laws too. Second, the full richness of finding a route involving human interests etc. might actually fundamental but in the aesthetic aspect for example, rather than application. The application level would be finding a route in Britain rather than in the USA, for example. ] []
# AB: Now, to encapsulate in a computer the laws of the spatial aspect (or the lingual aspect, or whatever it is) is actually quite a task. Task that makes us humble, because there is so much that don't understand. []
# Let's say [example], of the lingual aspect, so that an AI system that can produce Apps that are to do with language processing, /
# So: Sorry, I'm blethering. But there is two things. One is encapsulating the laws of each aspect, and my vision / and the other one is using those [software] modules that have the laws of each aspect, for building Apps. []
[1.00.15]
Contents
# AB: Now, my vision has long been "Can we produce a general purpose AI system in which the laws of each of the fifteen aspects is adequately captures?" [] ***
# Big long, multi-decade project, there! []
# JC: I'd devote my professional life to that, AB. [] ***
# I'm just saying that right now. I would 'd devote my professional life from this point forward. []
# AB: That's a computer science thing; it is, if you like, my computer science interest. Which was there until the late 1990s, then I had to choose to do more higher-level stuff of use and so on, and Dooyeweerd. []
Contents
# SJ: AB, can you express that again what you just said? How? What is that?
# AB: I'll express it slightly differently [going back to roots of AB's vision]. In the 1980s, when I was working at ICI, I had the idea that there was different aspects of knowledge. (And I did not know Dooyeweerd then.) []
# And each aspect of knowledge required a different set of algorithms to make it work properly. []
# So,
# And at the time, in the 1980s, different AI systems, or knowledge representation languages, were focused on different aspects. So there was a lot of / []
# So, have you got the idea of different aspects of knowledge? # SJ: Yeah.
# AB: And then there's the lingual aspect [AB: which actually at one point I added as the aspect of text], and the social aspect. - This is before I thought of, or knew, Dooyeweerd; I'd worked out these aspects. []
# Then I realised that the aspects I'd come up with intuitively were actually matched some of Dooyeweerd's aspects. And I was quite excited. []
# Because the issue is, if you try to represent knowledge of one / [knowledge] that's meaningful in one aspect, in another aspect, then you are likely to get errors in your computer program. []
[AB: Errors occur because we are trying to express the laws and properties of one aspect in the laws and properties of another, and it becomes very clumsy to do so at all, and impossible to do so fully. ]
# AB: So, let me give you an example, a simple example. There was a [computer] language [AB: or "knowledge representation formalism"] called PROLOG (Programming in Logic). And all its programs were:
Predicate / like
"X is the Uncle of Y, IF X is the Brother of Z, AND Z is the Parent of Y. [for some X,Y,Z]"
Yknow, logical statements like that. []
[1.03.45]
# AB: And, to PROLOG, these statements were purely / each word was purely a token. It did not know the idea of Uncle and so on, in order to, if you like, intelligently work out things. You would assemble hundreds of these statements, and that would be your program. And you would give it some data and it would give you an answer, basically. []
[AB: And the program would be like that, and when the program is run, the user would give names for X and Y and the program would return the result "TRUE" or "FALSE". It would also, of course, have some knowledge of parent-child relationships, such as found in a family tree.]
# But it could not do arithmetic. To it, the number 4 and the number 5 were purely tokens, and it did not understand that 4 was less than 5. Unless you actually told it that 4 was less than 5, and then, 4 less than 6, and 5 less than 6, [and so on]. Yknow, it was just pure tokens. It did not understand the idea of less-than and more-than, and it did not understand the idea of arithmetic. []
# So, what the PROLOG people had to do, in their compilers, was to have, if you like, special predicates that did stuff in the background. Yknow, so, when you said you can have a predicate
"X is Sum of Y, Z, A, B and C and so on",
it would [in the background] sum together the numerical values of X, Z, A, B, C [he meant Y!] and see if it they match X, the numerical value of X. [AB: Then it would return a truth-value result of TRUE or FALSE.] But it was doing that in the Assembler [machine] code in the guts of the [computer] engine. []
# And, if you tried to just do arithmetic with PROLOG, it was really clumsy. []
# And when you have a clumsy program, it is open to errors. []
# And, likewise [in reverse], to do items and relationships, entities and relationships and properties, using pure numbers was also clumsy and inappropriate.
# And there was this idea of inappropriateness.
# AB: Now, does that make a bit more sense?
# SJ: Yeah.
# AB: Sorry, it's just as long as the other answer. But it /
[AB: Actually, AB should then have reverted to SJ's question and ended with the implication from all that, that we need separate modules for arithmetic, entities-and-relationships, and logic, and so on, and got away from PROLOG. Then it would have made a bit more sense. ]
# AB: And so, the idea is, "Is it possible to capture the laws of all the aspects, each aspect in a separate kind of module, which you could then offer to developers?", to say, "Here we are; in this application, you need the Spatial module and the Quantitative module, and the Lingual module,, and the Pistic module." [] ***
[1.06.45]
Contents
# SJ: Does that include telling them that from each aspect you can reach out to the others? [] ***
# JC: SJ, that's my exact question. Where does it go in retrocipocity and dependency? Absolutely. []
# AB: [laughter] That's great! That's the weak point. In principle, yes, because each module would have to have / yknow like, the Social module aspect would need to access the Lingual module aspect. And that is inter-aspect dependency. []
# But, that is only a twinkle in the eye idea. Haven't actually built this thing yet. And I don't think I will now. []
[AB: Actually, I did, in a way. In his software, IRKit, developed during the 1990s, AB devised a system whereby modules could call upon each other in a defined order of dependency. ] []
# JC: What would you need to build it? # AB: A couple of centuries! []
# JC: [laughter] AI takes time away, so /
# AB: Before I understood Dooyeweerd, I did start building some of these modules. I've got an inference net program that's very powerful actually, more powerful than anything else. But it worked on the Amiga, so no-one used it. []
[AB: For the KBS software, see Istar. ]
# But it was the background ideas of it that were so important. Because of the background ideas / Och! nevermind about that! []
# It would need a big team. It's a huge research programme. And I don't know how to start, and I'm doing other things. But if anyone is interested in actually doing this, let's do it.
[AB: I think there are several questions that will require research. 1. What are the laws of each aspect? That has not been systematically discussed. 2. How to represent them as bit patterns and bit-operations in a computer. 3. How each aspect depends on others. 4. What input and output languages (graphical and text) would be appropriate to 'communicate' with such knowledge. 5 Data transfer protocols. Some of these are discussed in Chapter 7 of Foundations of Information Systems: Research and Practice. ]
# SJ: AB, I think is, if this A ??aim?? / I mean / Perhaps, in the future engineering stage of the machine learning pipeline, because that is the stage that we try to think in a modular way as well. # AB: Interesting. # SJ: So I think that perhaps could be a starting-point. [] ***
Contents
# JC: Can you two make sense of that to stupid people like me? Because I understand what you mean, but I can only see it in aspects. I understand when you say it, about inter-module communications in the background as the dependency and the retrociposity to check. But I have no idea where that goes, and the code development, within the compilers and the background, and stuff like that. []
# JC: Could you write something up and draw an actual 'this is what we are talking about' in a picture? # AB: Yeah. In fact, I have got stuff written up about it from the 1990s. []
# ACTION AB: Write an introduction to aspectual modules with a drawing.
# AB: I devised this knowledge representation kit, as it were. It wasn't built as Dooyeweerd's aspects, but it was kindof built as aspects. []
# What I did: I started with Assembler Language programming, and I thought "What the computer does, is to manipulate bit patterns. []
Contents
# AB: And so numbers (integers) is just a bit pattern interpreted in the normal binary way, where the first bit is 2 to the power 0, the second bit is 2 to the power 1, second [third] bit is 2 to the power 2, up to whatever wordlength it is, 2 to the power 64 or 32 or whatever.
[AB: Explanation: Each bit takes the value 0 or 1, and the quantitative meaning of, the quantitative amount represented by, any given bit pattern is the sum of all those powers-of-2 for which their bits are 1. For example, the bit pattern:
00001011would be 1 + 2 + 8 = 11. The right-most bit is bit-0, so represents 2-to-power-0, which is 1, and the left-most is bit-7, representing 2-to-power-7, which is 128. ] []
# And in most computers is already hardware to do addition and subtraction [and multiplication and division]. So that's, if you like, the quantitative aspect. []
# AB: But there are additional things [that are meaningful in the quantitative aspect] like - em, you don't just want integers, and real numbers (floating point numbers, continuous numbers), you want things like proportions, you want things like probabilities, you want things like ratios. []
# [Example] Yknow, you've got a ratio of 37/3 or something - usually we think of those as decimal numbers, but in terms of thinking of the laws of the quantitative aspect, there is the idea of ratio. How do you hold that in the computer as bit patterns / Well, obviously the answer is to have two integers side by side. []
# And then, how do you add two ratios? Well, that is ifyoulike, school arithmetic. A/B + C/D is whatever it is, (DA + BC) / BD. Whatever it is, I cannot remember. []
# And so, you would build into the quantitative module, not just the addition and subtraction [of integers, which is already in the hardware], but how add together and subtract ratios, how to multiply ratios, and so on. []
# AB: Then you've got things like prime numbers, yknow, there is a whole lot of stuff there about testing whether a number is prime. That's a fairly fundamental quantitative operation. Finding the next prime number: that is a huge operation when you've got large prime numbers. Yknow, as the Bitcoin people have found. []
# Got things like: If you are doing things with ratios, you need to factor out common things. Like 6/4 you have got to reduce it to 3/2. And there is a Euclid algorithm to do that. So you build that in. []
# AB: Got the idea? So it's not just adding and subtracting, it's everything to do with the quantitative [aspect] / with quantities.
[AB: The important thing here is that there is a host of concepts meaningful in the quantitative aspect and the basic quantitative operations on them work differently. And so all these have to be recognised and work efficiently in the module for the quantitative aspect. Types might include: simple quantities, ratios, probabilities, proportions, and all these would operate in different ways. And similarly for all other aspects. ] [] ***
[1.13.56]
Contents
# AB: Now, let us go to the spatial aspect.
# The spatial aspect is not just lengths and areas; it's things like / []
# I don't know whether you can see me on screen, but all I can see is the document. # JC: No. Can we see you?
# AB: But what I've got is a fist with the other hand surrounding it. Now, that idea of surrounding is a spatial, fundamental spatial idea. []
# It cannot easily be represented in numbers. The only way you can represent it in numbers is to, if you like, / []
# Let's say you've got a piece of woodland surrounding a field - whether it is fully surrounding or partially surrounding, it does not matter. You can only reason about that surroundingness by / if you use numbers, all you can do is express the boundary of the woodland in (x,y) terms, a list of (x.y) coordinates. []
Figure 1. Woodland surrounding fields (a), its expansions (b,c).
# And then if you say, for example, suppose there are foxes living in the woodland, and every night they come out and raid nests, birds nest that are on the ground in the field. Got the idea? # JC: Yes. # AB: And a fox will come out up to 200 m. So which bird nests are under threat from the foxes, and which birds nests are not? What you have got to do, to calculate that, you have got to take the shape of this field, which is like a kind of U shape or C shape or something, and expand it orthogonally by 200 m all round. And then you can find what nests are in that area. []
# But the algorithm for expansion is not simple. [AB: It is not a simple magnification; it is extending the boundary outwards.] If you have a boundary that is a C shape [(a)], then, if you expand it enough, then the expansions of the two ends of the C meet [(c)]. And now you've got a shape with a hole in it.
# How do you represent that in the computer? []
# Most people / most programmers who are programming an expansion algorithm would forget, not even think of the possibility, of creating a shape with a hole. []
[1.16.50]
# Can you envisage what I'm meaning: expand the shape.
# There is something that needs to be built into the spatial [aspectual] module. How do you do it? []
[AB: The usual way to represent a two-dimensional shape is as a list of (x,y) coordinates. Easy in a computer: you use an array of them. But the shape with a hole requires two such lists, not one. If we have not allowed for that, then we cannot process shapes with holes. And the expansion algorithm would give an error. ] []
[AB: So, like the quantitative aspect, we have many concepts in the spatial and many operations on them. Expansion is one, resizing is another, joining or splitting are two more, and so on. ] []
# AB: Then there's the kinematic module, physical module, biotic module, [and so on] []
Contents
# JC: I have a recommendation, may I recommend something? I don't know when this goes, but why don't we do an all-call for everybody to start to discuss that, AB. []
# And you and SJ are pros in this, and NB is pro in this. And I can be the resident, yknow, lay person. That, you speak in your necessary language, however that is the two of you cn understand the necessary fine-tuned points. And then we go, aspects to aspect to aspect to aspect, diligently, for two centuries! or yknow whatever, and just start to re / address that? []
# JC: SJ, does that sound interesting to you[] ***
[1.18.30]
# AB: What are you suggesting, JC, that we go through to the RLDG and say "Who would be interested?"? []
# JC: Yes, and I think this is where we start to / What I kindof envision is we, we understand our own little foibles and skills and weaknesses and assets that we bring. []
# And we develop both the parameters, that NB was talking about, both the social, societal and philosophical aspects of where, to SJ's point, can we really examine flourishing in a good way, compassion in probably the most explicit way. And examine that on one discussion thread, on one type of thing. []
# And we also have designated technical discussion, designated technical discussion upon each aspect. [] ***
# JC: We don't ??think encapsulate?? the time. It's only one time or half time or two times, we discuss it until the people who know this stuff, like you two, NB, and others, can say "That makes sense" and we always develop it with / []
# Understand that you/we need these compilers to work in the background and know how they interface. So you develop that, and that would be the inner working knowledge. []
# And the ethical assumption on the technical side is that you must have interactivity [between modules] at all points. []
# Don't let that be a hindrance but let that guide the modular development of each aspect. []
[1.20.00]
# AB: Great!
# AB: Are you suggesting, JC, that this is what this group does, over the next two centuries, or few months anyway? # JC: Yes. # AB: Right, yeah, great? []
# JC: Sounds good! # SJ: That would be good. Because it would be interesting. # JC: I think it would be / # SJ: Creative, innovative. []
# JC: As long as it makes your soul feel you are coming alive. This is where you have to check yourself, "My God, do you / is this how I steward my talents, abilities and time?" Then, yeah. []
# I mean, see what happens. # SJ: Yeah.
# AB: I've got a chapter of my book, which, to some extent, writes some of this up. So maybe I can extract the text from that and send it round. # JC: That's great. []
# ACTION AB: To extract text from book and send it round.
[Ed. In fact, AB sent whole chapters of both Philosophical Frameworks for Understanding Information Systems and Foundations of Information Systems: Research and Practice to JC and SJ, and can send to any others who would like them.]
Contents
# AB: What shall we do then, next time, in a month's time? Should I try to present this to a wider audience, or something? []
# JC: I think we call people and, yeah, present that framework. And kindof lay out and see if we can get buy-in and interest, and how they feel and they can be both appropriate and serve in this ministry, and move in that way. []
# AB: JC, would you like to / could you write a draft invitation to people? Because I can do the technical stuff, but I'm not quite sure how to invite people. So, if you can do a draft invitation to people. We can work on it, between the three of us. Send that to SJ and me. Have you got each others' emails? # JC: I'm sure I've got some??? SJ what's your email? # SJ: I'll type it in the chat for you. []
# ACTION JC: To draft an invitation to send to the RLDG. []
# JC: I invite people and connect with people. That's my job. []
# AB: That would be really good, if you could do that.
# AB: By the way, JC, I am / Some time ago, you suggested using Google Calendar for these meetings. I'm not averse to that. I just dunno how to do it, and so on. And so if we want to do that, then that would be great. # JC: I can be a kind of facilitator, administrative secretary. []
# AB: But the important thing is / Do you need to know everyone's emails in order to set that up? # JC: Yeah, I will eventually, everyone's emails, yeah. # AB: The problem is, I'm not allowed to tell / send you the emails because of / # JC: GDPR. # JC: Well just get their signoffs. Everyone's signoffs if they are OK. This is how we guage /
# The stage is "Are you OK to be included in this email from JC?" Like, if they say Yes to you, I can be given their email. That's their disclosure of like / [] ***
# AB: But most people don't reply. # JC: Then they are not interested. []
# JC: What would be good is, I'll do the invitation draft to the two of you. And then you AB send the invitation, OK. And say "This is going to introduce what we might pursue" and then we will work on two paths, societal considerations and compassion, and path of technical development across all aspects. And then what will happen is: The invitation, you just have to, like, "If you want to be part of these groups going forward, please is it OK for JC to email you?" And then I'll move. []
# AB: Yeah, OK, thanks.
# ACTION JC: To compile draft invitation. [Ed. Done]
# ACTION AB: To finalize and send out such an invitation. [Awaiting comments]
# AB: Well, will that do for today?
# SJ: Yeah.
# AB: I'm going to go back to Zoom. Have you saved the document, JC? # JC: Yeah. # AB: / # JC: I'll get it over to you, AB.
[Ed. Some technical discussion about getting the saved document]
# ACTION JC: To send document to AB. [Ed. Done]
# JC: [closed in prayer, including a prayer for data scientists]
# JC: Bye all. # SJ: Bye bye.
[1.27.00]
[AB then copied chats into a file.]
Contents
[1.27.35]