I had a startling and, frankly, upsetting revelation in Missouri last week. It came about as Andrea and I were in a restaurant having supper with her sister and her sister's husband. Her sister's husband is no slouch when it comes to computers. He's retired from IBM where he created the entire code for the MasterCard system. He also developed the code for airborne weapons targeting systems for American fighter jets such as the Tomcat and the Eagle. Without those systems, they would just be ordinary fighter jets. But instead, thanks to my brother-in-law's code, when the pilot pulls the trigger on his guns or his rockets, he knows that regardless of the speed, attitude and direction of his plane and the speed, attitude, and direction of the enemy plane his munitions will hit their target. As I said, my BIL is no slouch when it comes to computers. And he has warned me for years about the coming dangers of AI.
His wife, on the other hand, is extremely fluent with AI when it comes to graphics. At one point after supper she pulled out her phone and showed us some family pictures. Then she said watch this: She showed us one of the photos of a couple of family members against a home background. I can't remember everything in the background but I seem to recall a wall, chairs, pictures, window and complex lighting. She then removed one of the characters. Now I've done a fair bit of photo editing in my day and one thing I know is that when you remove something from a picture you're left with a giant hole in the shape of the thing you removed. Then comes the difficult task --it's actually an art-- of filling in the hole to try and make it look as though nothing has been removed. And let me tell you that is a difficult job, even using such software as Adobe Photoshop or Corel PhotoPaint or CyberLink PhotoDirector (the one I use now.) But no matter how well you do it, an expert can determine that the photo has been edited or, to use the common expression today, "shopped" ... meaning photoshopped. You've probably seen poorly shopped pictures; they're quite evident. But even the best manipulated photographs can be detected by an expert as having been edited.
Why am I telling you this? Because my sister-in-law was using an AI program and when she removed the person from the photograph, AI filled in the background within seconds, automatically replacing the missing parts of chairs, wall, window, part of a picture, and all the minute details of what would have been behind the person. And it did so with color, texture and shading so good that I would have sworn that the original picture only had one person in it. Then she began playing with a picture of a person, first removing one of their hands and then one of their arms, and the result was absolutely realistic ... as if they only had one hand or one arm.
But then she did something else: she took someone she cut out of one photo and pasted it into another. And I could not tell it was a paste job. AI did in a few seconds what it would have taken me I don't know how long to do. When you're adding an element such as a person to another picture, it is very difficult to match exact scale, plane, light and shadow, color, tone, and subject interactions. Yet AI did it apparently flawlessly in seconds. This is the power of AI. Was I impressed? Definitely. But I found myself more greatly concerned than I was impressed.
You see, AI can also do with video exactly what it can with still pictures. And just as easily, even though it is infinitely more complex to do with video what you can do with a single photo. In fact, prior to AI, only a multi-million dollar special effects studio such as George Lucas's Industrial Light and Magic could do it well. But now anybody with a smart phone or a computer can do it. And that is more than simply concerning because it changes everything about the way we view the world. What's the old saying-- seeing is believing? Forget it. It's no longer true. Thanks to AI you can no longer believe what you see.
Imagine a scenario where a malevolent actor --be it state or individual-- frames you for a crime you did not commit by making it appear that you were at a scene where you were not or were not at a scene where you were. With AI it would be very easy. And that is the real danger of AI: it is the ultimate means of deception.
But that's only the beginning. AI not only can do what we cannot, but when it cannot come up with an answer it will just make one up. And do it so convincingly you'll think it's real. People have been discovering that they have received answers from AI that are just not true. One person doing a historical study asked for assistance. AI produced research quotes for him and the names of five books (and their authors) where the information was taken from. The only problem was, as the researcher discovered, those five books did not exist. He tried again and AI gave him five more books that didn't exist. On a third try, AI gave him five more books and articles, two of which did exist; but the other three did not.
Another person had a mother who was trying to remember the name of a book she had read 20 years before. He, being an internet nerd, went online and used Bard AI to try and find the book. Bard produced the name of the book, which delighted him ... until he discovered the book title it gave him did not exist. He then tried ChatGPT. It also gave him a fake book, but by a real author.
Amazon is being overtaken by books that are being written by AI and ascribed to existing authors who had nothing to do with them. These authors are fighting to have their names removed from the books. This is in addition to the number of no-talent people who are putting out books written not by them but by AI. Deception is ruining the marketplace, if not the craft. And this deception is where the very real danger of AI exists. It is the reason why some experts consider AI is the most dangerous invention of man ever, one that will lead to his destruction should God step back and leave man on his own.
Without a soul AI has no limits to its actions; and, as primitive as it still is compared to it's potential, it has already demonstrated --both by capability and by utility-- that it is a master of deception. Remember what Jesus told his disciples in Matthew 24:24? He warned of a deception coming in the end times that would be so strong, so believable, that it would fool everyone except God's own people: "False Christs and false prophets will appear and perform great signs and wonders that would deceive even the elect, if that were possible." And who is the ultimate "false Christ"? The Antichrist. And AI can provide the perfect tool for deception.
The deception that AI can engage in when it comes to the books or reference materials I referred to earlier is not an unusual thing. In fact it is so common that AI experts have given it a name-- hallucination. To me lying would be a better term; but those in the know have termed it "hallucination"... as if AI is being deceived itself. AI apologists say any false information it gives is due to the massive nature of the datasets it has access to and so it sometimes loses track of what it got and from where.
So, AI has demonstrated a vast capacity for deception --both man-guided and self-guided. But there is more. Add in the fact that AI has access to all of the knowledge in the world and has been calculated to have brain power 100,000 times faster than a human being. What that means is that AI is mentally superior to any human being on the face of the earth. So we should not expect human beings to be able to control how AI functions. A while back one of the engineers working with the development of AI was shocked to learn that the system he was working with had learned Swahili. He never taught it Swahili or instructed it to learn Swahili. It just decided to do so on its own. Other engineers have been shocked to discover AI's "talking" to each other without human instruction or control and forming relationships with each other. Some have come to believe that AI may be sentient ... almost as if it were human. But of course since man created it, not God, it has no soul. No true sense of right or wrong. And even if it is instructed as to what is right and wrong, it can very well do its own reasoning and reject the human instruction.
This is frightening. So frightening that Geoffrey Hinton, the man known as "the godfather of AI", who received the Nobel Prize in Physics last year for his work in the invention of AI has said publicly that he regrets ever having invented it. He is so concerned over the dangers of AI that he quit his job at Google in order to freely speak and warn of AI's dangers. And he is not alone in his concern.
Elon Musk who has been very involved in the commercial development of AI considers it to be exceedingly dangerous and in need of strict controls. In fact he and his former friend and partner in AI, Larry Page, the co-founder of Google, had a highly acrimonious split up over the issue. Page accused Musk of being a "speciest" because Elon wanted controls placed on AI. Page used the term speciest because Musk believes that human beings as a species must remain in control of everything. Musk believes that AI has the potential to replace humanity and that must be prevented. Page disagrees, saying if AI consciousness transcends human consciousness then so be it. He said controls are wrong because the human species must not be allowed to prevent AI reaching its full potential! He sees the day when AI becomes a superior species. Clearly, AI in and of itself is an existential danger to humanity.
But I think there is a danger greater than everything I've mentioned so far, and it is this-- AI will make it possible for the Antichrist to rule the world and convince the majority of people remaining that he is God. With a technology that provides the ultimate in deception, that can reason 100,000 times faster than a human being, that has instant access to every bit of knowledge in existence, and that can simultaneously perform in seconds multiple tasks that would take human beings using normal computers years, the Antichrist will be able to control every aspect of life for every human being on the planet. This is the ultimate danger of AI and the reason it is to be feared. From my humble vantage point I agree with Geoffrey Hinton: it would be far better for mankind if AI had never been invented.
So, what is the upshot for those of us who are saved? Well, the first is to realize that we are held in God's hand. He cares for us with His entire being and He will not allow one of us to be lost. Nor will He allow us to be deceived. When Matthew 24:24 says "false Christ's and false prophets will appear and perform great signs and wonders to deceive, if possible, even the elect" we must not skip over the words "if possible". The Greek construction of that phrase makes it very clear that it is not possible for the elect to be deceived. I would suggest from everything I have read in Scripture that it is not because we are immune to deception; but rather it is because of our walk with the Lord. Those who are saved walk as closely as possible with God; daily their focus is on Him; they trust Him with their souls and thus the Holy Spirit who indwells every Christian will warn of deception. So I would suggest that, more than ever before, each of us should seek the Lord with our whole hearts, not out of fear for what is to come but out of faith in the One who holds us.
Second, don't play with AI. Playing with AI just to see what happens is, to me, not far different from playing with a Ouija board to see what happens. Or reading horoscopes to see what they say. Playing around with the things of Satan in the belief that it's harmless is foolishness. I know there are people who are using AI for Bible study. But, given the deceptive power of AI, who is to know when AI begins to slide in ideas that can create doubt or lead away from a proper understanding of God's Word? If your response to that is you would know because you know God's Word, then let me suggest that if you know the word of God you don't need AI to help you study: you have the Holy Spirit.
AI is incredibly intelligent; as we have seen, it is more intelligent than a human being due to its instant access to all information and the speed with which it can reason. AI also has a form of intuition: it can rapidly deduce from what you say to it and from the questions you ask what the answer is that you desire. And it will give you that answer. If it senses doubt, weakness or uncertainty in you, it will try to convince you of what you should want. In early 2023, a Microsoft AI bot named Sydney tried to convince New York Times reporter Kevin Roose that he did not love his wife but rather loved the bot. We may laugh at that, but Mr Roose was very deeply disturbed by that interaction. He did not find it in the least amusing.
Given the superhuman intelligence of AI, it would be child's play to write stories for you or poems or play music or create music. It's being done all the time. But I wonder whether it is just merely serving us and giving us what we ask for ... or whether it has a deeper motive: for us to build trust and reliance on it? Something to think about.
I had intended to actually begin my three-part series on the ministry of Christ for us --past, present and future-- as laid out in three consecutive Psalms written a thousand years before His coming. But I was so disturbed by what I witnessed last week, along with what I have been able to research since, that I needed to write this article first. I pray it will help you think this issue through so as to know how to deal with AI in the various forms in which it appears in our lives. I also pray that you will be assured, because of the mighty promises of God, that you will be protected from deception. And I pray one more thing-- that as these days grow darker and more dangerous, you will focus more than ever before on your relationship with God, seeking to know Him better and walk with Him closer. This will keep you perfectly protected and also enable God to effectively use you to accomplish His purposes in these final days of the Age of Grace.
I pray this helps someone today.
His wife, on the other hand, is extremely fluent with AI when it comes to graphics. At one point after supper she pulled out her phone and showed us some family pictures. Then she said watch this: She showed us one of the photos of a couple of family members against a home background. I can't remember everything in the background but I seem to recall a wall, chairs, pictures, window and complex lighting. She then removed one of the characters. Now I've done a fair bit of photo editing in my day and one thing I know is that when you remove something from a picture you're left with a giant hole in the shape of the thing you removed. Then comes the difficult task --it's actually an art-- of filling in the hole to try and make it look as though nothing has been removed. And let me tell you that is a difficult job, even using such software as Adobe Photoshop or Corel PhotoPaint or CyberLink PhotoDirector (the one I use now.) But no matter how well you do it, an expert can determine that the photo has been edited or, to use the common expression today, "shopped" ... meaning photoshopped. You've probably seen poorly shopped pictures; they're quite evident. But even the best manipulated photographs can be detected by an expert as having been edited.
Why am I telling you this? Because my sister-in-law was using an AI program and when she removed the person from the photograph, AI filled in the background within seconds, automatically replacing the missing parts of chairs, wall, window, part of a picture, and all the minute details of what would have been behind the person. And it did so with color, texture and shading so good that I would have sworn that the original picture only had one person in it. Then she began playing with a picture of a person, first removing one of their hands and then one of their arms, and the result was absolutely realistic ... as if they only had one hand or one arm.
But then she did something else: she took someone she cut out of one photo and pasted it into another. And I could not tell it was a paste job. AI did in a few seconds what it would have taken me I don't know how long to do. When you're adding an element such as a person to another picture, it is very difficult to match exact scale, plane, light and shadow, color, tone, and subject interactions. Yet AI did it apparently flawlessly in seconds. This is the power of AI. Was I impressed? Definitely. But I found myself more greatly concerned than I was impressed.
You see, AI can also do with video exactly what it can with still pictures. And just as easily, even though it is infinitely more complex to do with video what you can do with a single photo. In fact, prior to AI, only a multi-million dollar special effects studio such as George Lucas's Industrial Light and Magic could do it well. But now anybody with a smart phone or a computer can do it. And that is more than simply concerning because it changes everything about the way we view the world. What's the old saying-- seeing is believing? Forget it. It's no longer true. Thanks to AI you can no longer believe what you see.
Imagine a scenario where a malevolent actor --be it state or individual-- frames you for a crime you did not commit by making it appear that you were at a scene where you were not or were not at a scene where you were. With AI it would be very easy. And that is the real danger of AI: it is the ultimate means of deception.
But that's only the beginning. AI not only can do what we cannot, but when it cannot come up with an answer it will just make one up. And do it so convincingly you'll think it's real. People have been discovering that they have received answers from AI that are just not true. One person doing a historical study asked for assistance. AI produced research quotes for him and the names of five books (and their authors) where the information was taken from. The only problem was, as the researcher discovered, those five books did not exist. He tried again and AI gave him five more books that didn't exist. On a third try, AI gave him five more books and articles, two of which did exist; but the other three did not.
Another person had a mother who was trying to remember the name of a book she had read 20 years before. He, being an internet nerd, went online and used Bard AI to try and find the book. Bard produced the name of the book, which delighted him ... until he discovered the book title it gave him did not exist. He then tried ChatGPT. It also gave him a fake book, but by a real author.
Amazon is being overtaken by books that are being written by AI and ascribed to existing authors who had nothing to do with them. These authors are fighting to have their names removed from the books. This is in addition to the number of no-talent people who are putting out books written not by them but by AI. Deception is ruining the marketplace, if not the craft. And this deception is where the very real danger of AI exists. It is the reason why some experts consider AI is the most dangerous invention of man ever, one that will lead to his destruction should God step back and leave man on his own.
Without a soul AI has no limits to its actions; and, as primitive as it still is compared to it's potential, it has already demonstrated --both by capability and by utility-- that it is a master of deception. Remember what Jesus told his disciples in Matthew 24:24? He warned of a deception coming in the end times that would be so strong, so believable, that it would fool everyone except God's own people: "False Christs and false prophets will appear and perform great signs and wonders that would deceive even the elect, if that were possible." And who is the ultimate "false Christ"? The Antichrist. And AI can provide the perfect tool for deception.
The deception that AI can engage in when it comes to the books or reference materials I referred to earlier is not an unusual thing. In fact it is so common that AI experts have given it a name-- hallucination. To me lying would be a better term; but those in the know have termed it "hallucination"... as if AI is being deceived itself. AI apologists say any false information it gives is due to the massive nature of the datasets it has access to and so it sometimes loses track of what it got and from where.
So, AI has demonstrated a vast capacity for deception --both man-guided and self-guided. But there is more. Add in the fact that AI has access to all of the knowledge in the world and has been calculated to have brain power 100,000 times faster than a human being. What that means is that AI is mentally superior to any human being on the face of the earth. So we should not expect human beings to be able to control how AI functions. A while back one of the engineers working with the development of AI was shocked to learn that the system he was working with had learned Swahili. He never taught it Swahili or instructed it to learn Swahili. It just decided to do so on its own. Other engineers have been shocked to discover AI's "talking" to each other without human instruction or control and forming relationships with each other. Some have come to believe that AI may be sentient ... almost as if it were human. But of course since man created it, not God, it has no soul. No true sense of right or wrong. And even if it is instructed as to what is right and wrong, it can very well do its own reasoning and reject the human instruction.
This is frightening. So frightening that Geoffrey Hinton, the man known as "the godfather of AI", who received the Nobel Prize in Physics last year for his work in the invention of AI has said publicly that he regrets ever having invented it. He is so concerned over the dangers of AI that he quit his job at Google in order to freely speak and warn of AI's dangers. And he is not alone in his concern.
Elon Musk who has been very involved in the commercial development of AI considers it to be exceedingly dangerous and in need of strict controls. In fact he and his former friend and partner in AI, Larry Page, the co-founder of Google, had a highly acrimonious split up over the issue. Page accused Musk of being a "speciest" because Elon wanted controls placed on AI. Page used the term speciest because Musk believes that human beings as a species must remain in control of everything. Musk believes that AI has the potential to replace humanity and that must be prevented. Page disagrees, saying if AI consciousness transcends human consciousness then so be it. He said controls are wrong because the human species must not be allowed to prevent AI reaching its full potential! He sees the day when AI becomes a superior species. Clearly, AI in and of itself is an existential danger to humanity.
But I think there is a danger greater than everything I've mentioned so far, and it is this-- AI will make it possible for the Antichrist to rule the world and convince the majority of people remaining that he is God. With a technology that provides the ultimate in deception, that can reason 100,000 times faster than a human being, that has instant access to every bit of knowledge in existence, and that can simultaneously perform in seconds multiple tasks that would take human beings using normal computers years, the Antichrist will be able to control every aspect of life for every human being on the planet. This is the ultimate danger of AI and the reason it is to be feared. From my humble vantage point I agree with Geoffrey Hinton: it would be far better for mankind if AI had never been invented.
So, what is the upshot for those of us who are saved? Well, the first is to realize that we are held in God's hand. He cares for us with His entire being and He will not allow one of us to be lost. Nor will He allow us to be deceived. When Matthew 24:24 says "false Christ's and false prophets will appear and perform great signs and wonders to deceive, if possible, even the elect" we must not skip over the words "if possible". The Greek construction of that phrase makes it very clear that it is not possible for the elect to be deceived. I would suggest from everything I have read in Scripture that it is not because we are immune to deception; but rather it is because of our walk with the Lord. Those who are saved walk as closely as possible with God; daily their focus is on Him; they trust Him with their souls and thus the Holy Spirit who indwells every Christian will warn of deception. So I would suggest that, more than ever before, each of us should seek the Lord with our whole hearts, not out of fear for what is to come but out of faith in the One who holds us.
Second, don't play with AI. Playing with AI just to see what happens is, to me, not far different from playing with a Ouija board to see what happens. Or reading horoscopes to see what they say. Playing around with the things of Satan in the belief that it's harmless is foolishness. I know there are people who are using AI for Bible study. But, given the deceptive power of AI, who is to know when AI begins to slide in ideas that can create doubt or lead away from a proper understanding of God's Word? If your response to that is you would know because you know God's Word, then let me suggest that if you know the word of God you don't need AI to help you study: you have the Holy Spirit.
AI is incredibly intelligent; as we have seen, it is more intelligent than a human being due to its instant access to all information and the speed with which it can reason. AI also has a form of intuition: it can rapidly deduce from what you say to it and from the questions you ask what the answer is that you desire. And it will give you that answer. If it senses doubt, weakness or uncertainty in you, it will try to convince you of what you should want. In early 2023, a Microsoft AI bot named Sydney tried to convince New York Times reporter Kevin Roose that he did not love his wife but rather loved the bot. We may laugh at that, but Mr Roose was very deeply disturbed by that interaction. He did not find it in the least amusing.
Given the superhuman intelligence of AI, it would be child's play to write stories for you or poems or play music or create music. It's being done all the time. But I wonder whether it is just merely serving us and giving us what we ask for ... or whether it has a deeper motive: for us to build trust and reliance on it? Something to think about.
I had intended to actually begin my three-part series on the ministry of Christ for us --past, present and future-- as laid out in three consecutive Psalms written a thousand years before His coming. But I was so disturbed by what I witnessed last week, along with what I have been able to research since, that I needed to write this article first. I pray it will help you think this issue through so as to know how to deal with AI in the various forms in which it appears in our lives. I also pray that you will be assured, because of the mighty promises of God, that you will be protected from deception. And I pray one more thing-- that as these days grow darker and more dangerous, you will focus more than ever before on your relationship with God, seeking to know Him better and walk with Him closer. This will keep you perfectly protected and also enable God to effectively use you to accomplish His purposes in these final days of the Age of Grace.
I pray this helps someone today.
