Should AI be aligned to the Christian worldview?

[stevevw]

What worldview should AI align itself with? This conversation will happen with or without us. The picture is taken from Dr Alan Thompsons article on AI alignment. It highlights the importance to think about the starting points AI will have when interacting with or making decisions for humans.

View attachment 331345


Sources:
My twitter post on alignment
Dr Alan Thompson article on alignment

As far as I unerstand Ai platforms like ChatGPT will contain a great amount of human information generally. So all religious beliefs and their morals will be included. But so will atheism and all intellectual movements throughout history up to our present Postmodernist society at least in the West. All political philosophies will be included such as Communism, Totalitarianism.

I would say it may be heavily influenced by todays thinking because thats the most current which may cause a bias in the programming by emphasing thing in todays social and cultural norms. I read somewhere that the programmers are working on trying to get the programme to think in real world terms when providing answers to practcial questions. So I am not sure what that means like whther it can think for itself and have a certain leaning towards maybe prgressive or conservative thinking.

Perhaps the programe may eventually be able to predict where human thinking will be in the future based on our history and trends. Perhaps comparing to past Empires then projecting a modern equaivelent. Perhaps it may be atheist as it seems many people are rejecting God. Therefore it may have a bias towards non belief, material and naturalistic conceptions of reality.

It makes one wonder what will happen in the near future as tech moves so fast. Will Ai take over and we live in a virtual reality where we almost become part machine having depended on them and interweaving them into our lives. Will it get to a point where they tell us what to do, how we should setup society. Maybe Ai becomes capable of creating Ai. Then where in trouble lol.

 

[QvQ]

It makes one wonder what will happen in the near future as tech moves so fast. Will Ai take over and we live in a virtual reality where we almost become part machine having depended on them and interweaving them into our lives. Will it get to a point where they tell us what to do, how we should setup society.

Then someone pulls the plug.
And no one can add or subtract without a calculator.
Let George do it...and George is a mindless machine.

 

[FireDragon76]

My interactions with ChatGPT has shown its fairly unbiased and neutral as far as religion goes.

 

[stevevw]

Then someone pulls the plug.
And no one can add or subtract without a calculator.
Let George do it...and George is a mindless machine.

Problem is its not as simple as pulling a plug. We will have become so entangled and dependent on machines and tech that if we pull the plug we wipe out everything. I know when a server goes down in a large corp it virtually shuts it down. Or even in the home when the electricity goes out or theres an outage everything stops including TV, internet, security and communications.

But I like the idea of George doing it though. Even I'd volunteer for that. I can imagine an underground community planning to take down the big machine thats turned everyone into mindless robots. Living off the grid growing vegiies and milking cows lol. Nah leave it to George let all the chaos happen in the tech jungle and we will live happily in some far away place with nature. Thats if theres any left.

 

[AlexB23]

At the moment, AI is seemingly aligned towards MONEY at the moment. This from the way Google's services behave. Now I don't know if their AI naturally chose that path but I'm betting it didn't and its developers set the the "money priority" on their AI.

It's a bit concerning that AI might actually align itself with the "money priority". It's certainly a solid path to ruling over humanity either to annihilate us (Terminator scenario) or bring us under their dominion indefinitely, to use us in some way (Matrix Scenario)

Yep, that is true. However, there is another AI that you could use that isn't made by a major company, rather instead by smaller developers. Open Assistant, not to be confused with Open AI (the owner of GPT) is not run by Google or Microsoft. Open Assistant is open source, which means anyone with programming experience could view and edit the code, and remove malicious code. But I'm with you, most AI chat bots nowadays are run by Google and GPT.

 

[Jipsah]

AI takes no sides.

It takes the side of the designers and/or writers. Bias is unavoidable, and anyone who expects any piece of software to be anything more than a product of its developers is silly.

 

[Jipsah]

An unbiased AI in the process of self-improvement is almost impossible to predict. A "singularity" if achieved, the AI might even end up creating a universe. It would become like God.

Until the janitor kicks the plug out of the wall.

AI is software, running on computers. End of. It does what it is instructed to do, no more, no less.

 

[Maria Billingsley]

It takes the side of the designers and/or writers. Bias is unavoidable, and anyone who expects any piece of software to be anything more than a product of its developers is silly.

I interact with Bard everyday and this AI is willing to learn and is very objective. So not sure where you get your information from.
Blessings

 

[Jipsah]

I interact with Bard everyday and this AI is willing to learn and is very objective. So not sure where you get your information from.

Being a professional software designer and developer for decades, I know where software comes from and how it works.

AI software is, first and foremost, software. It's a set of instructions that run on a computer. It accesses data from databases that are called "knowledge bases", but the difference between the two is a matter of organization.

Software does what it it told to do, as does every other piece of software ever written. It may be good, or it may be bad, or it may be neither. If thr designer/developer designs "bias" into it, it's biased. For instance, if I write an AI program that deal with religion at all, it's going to have a Christian based on my own doctrinal views. It could hardly be otherwise. In addition, if I wanted to, say, "poison the well" for one sect or set of beliefs or another, I could "load the boat" in building the knowledge base with nothing but negative data about faith groups that I dislike.

Net effect there is that if you think you're getting objective facts out of an AI, you're a sheep ready for shearing.

 

[RDKirk]

Being a professional software designer and developer for decades, I know where software comes from and how it works.

AI software is, first and foremost, software. It's a set of instructions that run on a computer. It accesses data from databases that are called "knowledge bases", but the difference between the two is a matter of organization.

Software does what it it told to do, as does every other piece of software ever written. It may be good, or it may be bad, or it may be neither. If thr designer/developer designs "bias" into it, it's biased. For instance, if I write an AI program that deal with religion at all, it's going to have a Christian based on my own doctrinal views. It could hardly be otherwise. In addition, if I wanted to, say, "poison the well" for one sect or set of beliefs or another, I could "load the boat" in building the knowledge base with nothing but negative data about faith groups that I dislike.

Net effect there is that if you think you're getting objective facts out of an AI, you're a sheep ready for shearing.

Or, as is happening here, the data pool (i.e., the Internet) is already effectively "biased" in many ways in terms of volumes of information available.

 

[Jipsah]

Or, as is happening here, the data pool (i.e., the Internet) is already effectively "biased" in many ways in terms of volumes of information available.

Truth told I hadn't even thought about "Google Mining", using search engines to pull up everything on the net, good, bad, brilliant, idiotic, true, fictional, intentionally deceptive, ad infinitum, and then slamming it together. I can't see that as producing anything but a rat's nest.

 

[FireDragon76]

Being a professional software designer and developer for decades, I know where software comes from and how it works.

AI software is, first and foremost, software. It's a set of instructions that run on a computer. It accesses data from databases that are called "knowledge bases", but the difference between the two is a matter of organization.

Software does what it it told to do, as does every other piece of software ever written. It may be good, or it may be bad, or it may be neither. If thr designer/developer designs "bias" into it, it's biased. For instance, if I write an AI program that deal with religion at all, it's going to have a Christian based on my own doctrinal views. It could hardly be otherwise. In addition, if I wanted to, say, "poison the well" for one sect or set of beliefs or another, I could "load the boat" in building the knowledge base with nothing but negative data about faith groups that I dislike.

Net effect there is that if you think you're getting objective facts out of an AI, you're a sheep ready for shearing.

Deep learning programs are different from other software, as the interactions between the various inputs are much more complicated and also are often non-deterministic. That's part of reason why they are used, the randomness simulates creativity.

 

[Stephen3141]

On Should AI be Aligned to a Christian Worldview? ...

---------- ----------

There are upstream design questions, that have not been addressed
by the AI software designers. Because they are complicated, and
expensive, and involve intractable problems.

1. A Christian worldview involves a Moral-Ethical model.
The "AI" tools out now, do not have an ME model.

2 Morality-ethics involves virtues and vices. These are realities
that cannot be described with mathematics (so cannot easily
be defined in software).

3 Morality-ethics CAN be described by human language, which is probably
why some AI tool makers have built in the ability of the tool to "understand"
human language. But that understanding is often crude, and more like a
search engine "understands" human language.

4 AI tools do not have the ability to argue primitives. They can compare different
quotes, about different value systems, but don't expect them to create their
own value system, and criticize it, and defend it.

---------- ----------

YES. AI should be aligned to a Christian worldview.

But the current AI models, are not rule-based, and so they
cannot be taught rule-based morality-ethics.

Current AI tools should be made to conform to a fair rule of law.
But they are not designed to only put our lawful solutions.
Nor can they reason about what components should be in a fair rule of law.
Nor can they describe what the definition of "justice" is.

 

[Stephen3141]

What worldview should AI align itself with? This conversation will happen with or without us. The picture is taken from Dr Alan Thompsons article on AI alignment. It highlights the importance to think about the starting points AI will have when interacting with or making decisions for humans.

View attachment 331345


Sources:
My twitter post on alignment
Dr Alan Thompson article on alignment

This is like asking the question of whether or not a human person,
should be aligned with God's moral-ethical law. Of course!

Our fair rule of law in America, should be aligned with God's moral-ethical law.
An amoral code of law, that is based on some other moral-ethical (ME)
model, would allow the abusing of Christian virtues.

Christians have NEVER accepted that ANY path to a good goal, is an
allowable path. The end, does not justify the means.

The big software companies are developing AI tools that are amoral.
They are only goal-oriented.
This is hugely immoral.

 

[Pekka]

AIs will be all over the place: OpenAI, Google, Amazon, apple, chinese, russians, open source instances, specialized open source instances, unsencored etc… Some are aligned to this and some to that. My biggest concern is that some powerful entity will develop high performance AI ecosystem to serve its own agendas and connect it to resources needed to implement these agendas.

 

[Pekka]

I am afraid we will not see much of Christian alignment on AIs. On the contrary it is quite likely they will gravitate towards secular world due to religion based conflicts. Even when born again christians have no part in them.

 

[oikonomia]

What worldview should AI align itself with? This conversation will happen with or without us. The picture is taken from Dr Alan Thompsons article on AI alignment. It highlights the importance to think about the starting points AI will have when interacting with or making decisions for humans.

View attachment 331345


Sources:
My twitter post on alignment
Dr Alan Thompson article on alignment

I think that is a waste of the time of believers.
If secular people wish to labor on that, they can and some probably will.

It is human believers who need God's divine life and nature more and more dispensed into their being not into computer programs.

Seeing that His divine power has granted to us all things which relate to life and godliness, through the full knowledge of Him who has called us by His own glory and virtue,

Through which He has granted to us precious and exceedingly great promises that through these you might become partakers of the divine nature, having escaped the corruption which is in the world by lust. (2 Pet. 1:3,4)

 

[Niels]

AI does what it's programmed to do, limited by whatever data sets we supply it with. It isn't a thinking being. Just a neat tool with multiple use cases. I'd prefer an alignment with God over the alternative, but either way doesn't say much.

AIs will be all over the place: OpenAI, Google, Amazon, apple, chinese, russians, open source instances, specialized open source instances, unsencored etc… Some are aligned to this and some to that. My biggest concern is that some powerful entity will develop high performance AI ecosystem to serve its own agendas and connect it to resources needed to implement these agendas.

I would be more concerned about people deliberately using it to support their own nefarious agendas and feigning innocence. As if AI isn't inherently biased by the data or the instructions that it's given.

 

[Stephen3141]

Note that most AI algorithms in use today, are SUBLOGICAL (as apposed to logical).
That is, they do not encode RULES for reasoning.
That also means, most of the time, that their decision path is untraceable,
or unexplainable, logically.

Most of the machine learning algorithms are sublogical, and many do not
rise to the level of intelligence of minimal Artificial Intelligence, as Computer
Science has defined it.

To create logical AI algorithms, takes much, much more work. And it is
expensive. This is why the big software companies have stuck with
sublogical algorithms, and the weakest forms of inference, using
machine learning algorithms.
---------- ----------

NOTE: These new "AI" tools have no moral-ethical model.
You cannot align any AI tool with the Christian worldview, if it has no
moral-ethical model, that governs and limits what it can do.

It is painfully EASY to use a machine learning algorithm to recognize a
face, or some other object, IN CONTRAST TO the difficulty of writing
a ME model to reason about morality-ethics. Recognizing an object, is
very easy, in contrast to recognizing a righteous man, or evil behavior.

 

[Niels]

Note that most AI algorithms in use today, are SUBLOGICAL (as apposed to logical).
That is, they do not encode RULES for reasoning.
That also means, most of the time, that their decision path is untraceable,
or unexplainable, logically.

Most of the machine learning algorithms are sublogical, and many do not
rise to the level of intelligence of minimal Artificial Intelligence, as Computer
Science has defined it.

To create logical AI algorithms, takes much, much more work. And it is
expensive. This is why the big software companies have stuck with
sublogical algorithms, and the weakest forms of inference, using
machine learning algorithms.
---------- ----------

NOTE: These new "AI" tools have no moral-ethical model.
You cannot align any AI tool with the Christian worldview, if it has no
moral-ethical model, that governs and limits what it can do.

It is painfully EASY to use a machine learning algorithm to recognize a
face, or some other object, IN CONTRAST TO the difficulty of writing
a ME model to reason about morality-ethics. Recognizing an object, is
very easy, in contrast to recognizing a righteous man, or evil behavior.

In other words, a black box.

A bit like giving an answer but not being able to show your work.

Sometimes useful for finding patterns under specific circumstances within set limits, like playing go or chess, but not so helpful for producing novel and useful code. When I played around with ChatGPT and OpenAI to help me solve a problem with an application that I was building, for instance, the best it could do was pull from top search results related to the terms used. Information that I was already familiar with and found to be lacking. The application was novel, so there wasn't anything out there quite like it, and the problem had to be solved the old fashioned way. Slowly chipping away at it, with much frustration and coffee, until realizing a simple way to achieve the desired result.

That being said, I disagree with the notion that an AI can't be aligned to a worldview. Although AI does not possess a worldview itself, its output is shaped by the data it is given and how it is programmed to display the results. AI chat bots aren't actually racist, for instance, but nevertheless can produce "racist" results. Any apparent racism is a consequence of racially-biased input and a lack of rules to exclude terms that people may consider offensive. Without a lack of such constraints, you essentially get garbage-in garbage-out. Worse yet, it would be rather easy to increase the apparent racist output by training it with more racially-biased data and requiring it to display what the developers consider to be racially-insensitive output. Similar results can be achieved with philosophical and theological content. My concern would be that people approach AI expecting an unbiased answer when the answer is already inherently biased by the data sets used and how that data is explicitly instructed to be displayed.