Churches and AI

 
Lb921
 
Avatar
 
 
Lb921
Total Posts:  4
Joined  29-06-2017
 
 
 
15 September 2017 07:24
 

What risk do people see that churches (or other organized religious groups) will start to use AI to target non-believers?

It seem likely algorithms will only get better, especially when it comes to identifying our psychological profiles, and big data systems can already predict many aspects of our psyche better than we can. AI systems, trained to deliver advertising, are adept at targeting messaging to users.

What then is the likelihood that such systems could be used to convert people to a religion by delivering content calculated to appeal most to them?

I can see no reason why a well-funded religious group would not want to use machine learning to target people. With such a system religious groups could effortlessly target atheists, agnostics or even followers of other religions simply by scaling up their server power.

New religious “startups” could even form that are based on growing a “user base” of believers by delivering “religious content” that appeals most to the deep psychology of users.
 
I wrote a longer article on the subject and would be keen to hear anyone else’s views on this topic
https://medium.com/@LochlanBloom/the-coming-battle-ai-extremism-and-the-new-war-of-ideas-a98689499b7c

 
Kalessin
 
Avatar
 
 
Kalessin
Total Posts:  128
Joined  02-07-2017
 
 
 
15 September 2017 08:56
 

The first thing is not to over-mistify and see AI as something completely opaque yet frighteningly perfect.  It’s not any of these things.(at least yet). 

What AI can do is automate and instantiate a range of persuasion triggers and attention mechanisms which then have a collective effect on individual and particularly group behaviour.  This is primarily an aspect of marketing - or at least that is the area where investment and the analysis is taking place - but of course it has wider implications and applications, as marketing (or persuading) is ubiquitous. 

Could churches (or the church in general) become a lot better at marketing - recruiting and retaining new adherents, entrenching and deepening existing members, maximising financial value from its members and also leveraging the value of the collective data?

Yes, and of course they/it are probably trying to do that as much as possible now.  In fact the Abrahamic churches have been notably successful at this for a very long time.  The effect of entering a wonderful cathedral, sharing a meaningful communal experience with others, seeing others who you respect and admire publicly sharing your worldview, imbibing a message that requires engagement but rewards you with feelings of fulfilment, and the exponential benefit you feel the more you deepen you commitment ... this is the dynamic that algorithms are there to support.

The design of a cathedral and the structure of a sermon can be considered analogous to an algorithm, in conjunction with the ability of a pastor to ‘cold read’ his congregation and make good judgements about the tone, frequency and theme of his messages, and so on.

So what can churches do better?  The obvious path (as for all marketers) is to copy Facebook, Google, Snapchat and so on.  The primary pay-off for the intense persuasion algorithms is more to do with monetising the data pool to advertisers, although the current processes have an inherent tendency to produce echo chambers which can grow and collapse ‘virally’.  However, the church could use social influencing models (as some charities try to do) to generate engagement, affinity and direct financial support - I am sure it already does.

With that in mind, there is some reasonably good big data which shows that getting really ‘sticky’ adherence from large numbers is quite difficult even with agressive AI persuasion models.  Facebook is a platform and people don’t feel emotionally bonded with it as a “thing in itself” in the same way they might do with, say, their football club or indeed their church.  The evidence from cause-related marketing is much more mixed than simply the numbers you get if you use platforms like Snapchat, Instagram and so on, so there is a difference between how AI can perform depending on the context.

So the simple answer is yes, the church can, is, does and should use AI as part of marketing, just as it has always used the best persuasion tools available throughout history.  However, the implications are not more sinister either because it’s the church or because it is using AI.  It’s a crowded, volatile marketplace and just ‘having’ algorithms or AI doesn’t guarantee you anything.

Targeting ‘non-believers’ is, in marketing terms, the same as trying to get people to switch brands or start using a new product.  How easy is that to do?  See lots of examples these days - do you not think ALL the big players are not using AI to make it happen?  And this is quite a big switch if you want it to stick and be real.  Perhaps something life, for example, getting people who currently don’t have or want guns to buy a gun (or the other way round).  Or getting people to become vegetarian ...not impossible, and AI can play a role, but I hope that puts your mind at ease a bit smile

This is within my professional area so happy to stay on this if you have Qs or views ...
Kalessin

[ Edited: 15 September 2017 09:00 by Kalessin]
 
Cheshire Cat
 
Avatar
 
 
Cheshire Cat
Total Posts:  947
Joined  01-11-2014
 
 
 
15 September 2017 16:02
 

If 12 years of Catholic School couldn’t make me a believer, then I’m pretty certain AI can’t either.

 
 
Lb921
 
Avatar
 
 
Lb921
Total Posts:  4
Joined  29-06-2017
 
 
 
18 September 2017 04:47
 

You make some very good points. I agree it is important not to over-mystify AI and what it is capable of at present as compared to theoretical limits for these tools and
I guess my points are more in the mode of a thought experiment.

I feel if any of these changes happen they will probably happen slowly, in an iterative way, rather some overnight takeover and my point is more aimed at where these things could end up as opposed to any dire warning or fear in the present.

Having said that I see significant difference in using AI to market products as opposed to spreading religious ideology.

As you mention people don’t feel “emotionally bonded” with a lot of the marketing from commercial entities or platforms such as facebook in the same way they do with their football club or church and I would absolutely agree. You then go on to say “Targeting ‘non-believers’ is, in marketing terms, the same as trying to get people to switch brands or start using a new product.” but I would argue that this is not necessarily true.

Even the most fervent follower of a brand is aware that it is a commercial construct created by a company and this in turn is simply a legal structure created by the company’s founders. They may identify with the brand on many levels but surely very few people believe that the brand is “real” in the same way that religious followers might believe god is real and part of their life. 

It may be that this is not an issue but potentially there is a fundamental difference here and one which can be exacerbated by AI systems. Once a person feels that a truth has been revealed to them, reads or watches something that resonates so deeply with them that they change their outlook and really believes something to be real that is a lot harder to shake than allegiance to a brand.

This is even more the case for children and while I doubt anything will happen overnight, an iterative AI system could easily start to identify and manipulate a generation of younger users towards such content tat might have a significant impact on their ontological perspective. If such systems were able to become adept at this it could profoundly change the outlook of people and these sort of experiences will then become much harder to change than someone switching from coke to pepsi.

 

Kalessin - 15 September 2017 08:56 AM

The first thing is not to over-mistify and see AI as something completely opaque yet frighteningly perfect.  It’s not any of these things.(at least yet). 

What AI can do is automate and instantiate a range of persuasion triggers and attention mechanisms which then have a collective effect on individual and particularly group behaviour.  This is primarily an aspect of marketing - or at least that is the area where investment and the analysis is taking place - but of course it has wider implications and applications, as marketing (or persuading) is ubiquitous. 

Could churches (or the church in general) become a lot better at marketing - recruiting and retaining new adherents, entrenching and deepening existing members, maximising financial value from its members and also leveraging the value of the collective data?

Yes, and of course they/it are probably trying to do that as much as possible now.  In fact the Abrahamic churches have been notably successful at this for a very long time.  The effect of entering a wonderful cathedral, sharing a meaningful communal experience with others, seeing others who you respect and admire publicly sharing your worldview, imbibing a message that requires engagement but rewards you with feelings of fulfilment, and the exponential benefit you feel the more you deepen you commitment ... this is the dynamic that algorithms are there to support.

The design of a cathedral and the structure of a sermon can be considered analogous to an algorithm, in conjunction with the ability of a pastor to ‘cold read’ his congregation and make good judgements about the tone, frequency and theme of his messages, and so on.

So what can churches do better?  The obvious path (as for all marketers) is to copy Facebook, Google, Snapchat and so on.  The primary pay-off for the intense persuasion algorithms is more to do with monetising the data pool to advertisers, although the current processes have an inherent tendency to produce echo chambers which can grow and collapse ‘virally’.  However, the church could use social influencing models (as some charities try to do) to generate engagement, affinity and direct financial support - I am sure it already does.

With that in mind, there is some reasonably good big data which shows that getting really ‘sticky’ adherence from large numbers is quite difficult even with agressive AI persuasion models.  Facebook is a platform and people don’t feel emotionally bonded with it as a “thing in itself” in the same way they might do with, say, their football club or indeed their church.  The evidence from cause-related marketing is much more mixed than simply the numbers you get if you use platforms like Snapchat, Instagram and so on, so there is a difference between how AI can perform depending on the context.

So the simple answer is yes, the church can, is, does and should use AI as part of marketing, just as it has always used the best persuasion tools available throughout history.  However, the implications are not more sinister either because it’s the church or because it is using AI.  It’s a crowded, volatile marketplace and just ‘having’ algorithms or AI doesn’t guarantee you anything.

Targeting ‘non-believers’ is, in marketing terms, the same as trying to get people to switch brands or start using a new product.  How easy is that to do?  See lots of examples these days - do you not think ALL the big players are not using AI to make it happen?  And this is quite a big switch if you want it to stick and be real.  Perhaps something life, for example, getting people who currently don’t have or want guns to buy a gun (or the other way round).  Or getting people to become vegetarian ...not impossible, and AI can play a role, but I hope that puts your mind at ease a bit smile

This is within my professional area so happy to stay on this if you have Qs or views ...
Kalessin

 
EN
 
Avatar
 
 
EN
Total Posts:  20303
Joined  11-03-2007
 
 
 
18 September 2017 08:45
 
Cheshire Cat - 15 September 2017 04:02 PM

If 12 years of Catholic School couldn’t make me a believer, then I’m pretty certain AI can’t either.

That’s what I was thinking.  If you’ve already rejected the message, what information could possibly be conveyed trough AI that would make a difference?

 
Lb921
 
Avatar
 
 
Lb921
Total Posts:  4
Joined  29-06-2017
 
 
 
18 September 2017 09:16
 

I agree it could definitely be harder for any AI system to convert an ex-believer. In fact someone who has been through 12 years of school and already rejected the message may be highly resilient to being ‘reconverted’

I think there are two important points though -
1) the teachers at any school today are only human and do not have anywhere near a full psychological understanding of each child. If they had been able to understand your preferences/behaviour better than you yourself did (as it is shown some AI systems can now do)  then the outcome of your schooling might have ben quite different.

2) it is not necessarily a case of presenting ‘information’ in the form of any logical argument, it is about steering your entire process of learning about the world. If instead of finding the catholic education ridiculous it deeply resonated with your psyche then it would be that much harder to reject. An AI system would not have to do anything as blatant as bombard a user with evangelical texts, it could subtly shift the perception of a child over the course of many years, playing on emotional response so that they identified more strongly with core messages of a given religion and manipulating fears and weaknesses to reduce the likelihood of rejection.

Neither of these are likely to be perfect but could greatly affect the “Failure rate” in a given school

EN - 18 September 2017 08:45 AM
Cheshire Cat - 15 September 2017 04:02 PM

If 12 years of Catholic School couldn’t make me a believer, then I’m pretty certain AI can’t either.

That’s what I was thinking.  If you’ve already rejected the message, what information could possibly be conveyed trough AI that would make a difference?