AI Has Nothing to do With the Truth

By Tom KraMer

Today it seems that everyone uses AI as a source of truth. And admittedly, it is easy to do so. Why? Convenience. When you need an answer to something, why not get that answer immediately? Then your can get on to more important things. However, the answers, descriptions, histories, and information we get from AI might be convenient for helping is make a decision to move in a certain direction, but if we are concerned truth guiding our actions we have to be aware that the information we get from AI has absolutely nothing to do with the truth. In fact, it is impossible for it to actually know truth.

What Is Truth?

We could wax philosophical here as folks have done for many centuries, and debate what the actual definition of truth is. But because we are Christians, we understand God is the creator of everything, and therefore the sole source of all truth. That means everything that God and Jesus Christ (God in human form) have said is truth.

There is another kind of truth though. If I forgot to tie my shoe today, and I later told you that I forgot to tie my shoe, I would have told you something that was truth. I could have also told you something else that was not true. I have the option to tell truth or not to. This is because I am human. God does not have that option, He is only Truth, Light, and Goodness. He designed it so I have a choice. This is the weakness of AI.

You see, the AI that most of us are used to operates on large language models (LLM) which basically means that it quickly scans all of the internet, finds all the information it can related to our query, and delivers it to us in the form of sentences. The problem here is, as we all know too well, is that anybody can write anything and put it on the internet. This means that AI has to somehow filter "good" information from "bad" information when it gives us answers. How does it do that? One way is to look for sources that are affiliated with subject matter experts on the topic at hand.

Here's the rub: what if those subject matter experts are misinformed? What if their writings are based on a world view that negates a universal source of truth? What if they have been trained by powerful lobby's with an agenda of financial self promotion? Can AI figure that out? Sadly, no.

Subject matter experts can be divided, and one side can accuse the other of misinformation. If the accused side gets demoted, fired, de-licensed, and de-credentialed because the other side has more political power and social influence, AI will use that information to rule out the validity of the information presented in the LLM by the accused side. AI doesn't care about this. It just tries to present the most accurate information based on what it finds in the LLM.

Here are some of the major problems with the reliability of AI:

• Hallucinations: This is false information given out as truth, which happens because the AI is trained to predict the next word, not to verify truth.

• Lack of True Understanding: AI "predicts", it doesn't "know" anything. It doesn't understand facts, it just predicts statistical patterns in language, with lack of reasoning.

• Bias and Training Data Limitations: If the data that it is looking at is biased or incomplete AI answers will be systemically misleading.

• Overconfidence: AI models are bad at saying "I don't know". They give definitive answers even when they are unsure, and users don't know when this is.

• Fragility: Small changes in input can produce very different answers. They can even contradict themselves over similar queries, giving different "facts".

  
  

These facts alone should give us pause as we think about how often we take output from AI as fact.

Jack Nicholson in the box office hit, A Few Good Men.

Reliance Can Become an Idol

The question we have to ask ourselves is what are we going to use as our source of truth? Just look at the huge reliance on AI for information today. It's pretty convenient. It can be very helpful if we are just looking for information, and we take the results with a grain of salt.

But that is a slippery slope. We can become more and more reliant on it until we believe everything we are getting from it is truth. When we use it as our source of truth, it becomes our god. That is very dangerous. Revelation 13:5 says "And it was given to him to give breath to the image of the beast, so that the image of the beast would even speak and cause as many as do not worship the image of the beast to be killed." (LSB Version)

Sound scarily familiar? Wow. I didn't see that coming in the 1970"s when I first read that verse. I had no idea what it meant. Now the idea of "givng breath" to something that wasn't alive and having that thing actually talk (when it seemed that it was unusual for it to talk when written) hits pretty close to home. The sad thing is that this verse shows that many people "worshipped" That beast. Is AI the "Beast" of the book of revelation? Who knows, but I challenge you to find anything that looks like a better fit.

Remember, the book of Revelation tells us the Beast is controlled by Satan, and a very popular political figure of that time. We know from the words of Jesus that the goal of Satan is to steal, kill, and destroy. This should give you pause and turn up your alertness level to an 11 when it comes to AI.

Supposed Truths Written By Fools Are Not Wisdom

The covid-19 situation made the whole world a stage for the censorship of information. In past times scientists welcomed debate on scientific issues, conducted experiments, discussed them, and came to various conclusions about information. During the covid-19 debacle though, this did not happen. Anybody that presented information that did not match the information that the world powers wanted out there was not discussed, but was systematically discarded, and those who presented the information were mocked, ostracized, rejected, censored, and persecuted with the loss of their jobs and professional positions. This means only one source of information was being counted as "true" by people that were hailed in the media as experts, while those whose information was discarded were regarded in the media as purporters of "misinformation", and the weight of the media writings skews the information in the LLM in a way that AI is sure to pick up on it.

A fool writing to put his thoughts into the world.

Things To Remember

So what do we do? Go build a cabin in the woods and survive off gardens and squirrel meat? As attractive as that may sound ;) it might be better to realize what you should be using AI for and what you shouldn't. Think about these things when you consider using AI:

1. Know when you need truth: If you actually need truth in your current query, take your AI results with a grain of salt and research them the old fashioned way. Find other answers and opinions. Always make sure to include answers from people with a Christian worldview.

2. Don't make it an idol: God should be your resource for knowing what you should do in life, and when you need spiritual, philosophical truths. Remember, revelation paints a dismal picture for those who worship the beast.

3. Understand its source: Remember, any fool can write something and put it on the internet. If a lot of people do that in a similar way, you have a body of data that AI will be looking at.

Be Wise

Hopefully this helps you think a bit deeper about your use of, and the future of, AI. Keep these things in mind as the years roll by, and AI becomes a more embedded part of everyone's life. It just may help save yours in the end.

Thank you for joining me on this interesting exploration. Here’s to your health, your spirit, and the abundant life ahead.