The New Buzz Words: Artificial Intelligence

by | Feb 2, 2024 | Features, Times of Restoration

While TIMES OF RESTORATION does not take political positions, we feel it important in this column, in order to provide basis for further understanding and prayer, to discuss selected current events and trends in the light of God’s unchanging Word.

In this issue, we solicited computer programmer and chess player Ken Jones of Nova Scotia to investigate the emerging phenomenon of Artificial Intelligence and how Christians might be affected. His even-handed coverage will extend into the next issue as well.

Part I: And the whole earth marveled as they followed the beast (Revelation 13:3, ESV).

It seems like every time I turn around I see comments about Artificial Intelligence [AI] being used by some company or project. Just recently I was viewing a science video on YouTube, and part way through the 20-minute presentation, I realized that the speaker was flawlessly reading the script without pause. Sentences followed one another, but there was not time to get a breath between each one. Then it dawned on me that this perfect voice must be an artificial one. Sure enough, it was an AI engine that could read English text and speak it perfectly, with all the inflections in the right places.

Having used ChatGPT (the chatBot used by Microsoft’s Bing browser) several times to write computer code, I decided to ask it for a brief historical summary of AI. In order to do that, the chatBot uses an engine, GPT-4 [Generative Pre-Trained Transformer], which has to understand my request, look up the information from multiple sources on the internet, piece it together into a coherent structure, and present its findings to me in proper English. Here is what ChatGPT came up with in 2-3 seconds:

Artificial intelligence (AI) has its origins in the mid-20th century. Computer scientists started to create algorithms and software that could carry out tasks that ordinarily need human intelligence, like problem solving, pattern recognition and judgement. In 1956 the Dartmouth conference was held, which is considered the birthplace of AI. In the 1960s and 1970s, AI research was focused on problem solving and symbolic methods. In the 1980s, AI research shifted to machine learning and expert systems. In the 1990s and 2000s, AI research focused on data mining, natural language processing, and  robotics. In modern day AI, deep learning and neural networks are the focus. The future of AI is still being explored.

Now I need to explain what an algorithm is. From Wikipedia: “In mathematics and computer science, an algorithm is a finite sequence of rigorous instructions, typically used to solve a class of specific problems or to perform a computation.” A pocket calculator executes various algorithms depending on which key is pushed. An example would be the square root key, which executes an algorithm called Newton’s method. Another algorithm that everyone learns in school is long division, with its specified steps toward the desired result. 

Recently, since about 2012, AI has become very good at tasks which were impossible only a few years before. These include facial recognition, language tasks such as text-to-speech, voice recognition (in modern cars you can tell it to call someone on your cellphone while driving), translating from one language to another, accurate weather forecasting, creativity in artistic domains (painting, writing, photography etc.), and expert analysis of medical situations with multiple symptoms. These capabilities have led some people to think that AI is eventually going to take over the world.

In the near future ChatGPT will be getting a new engine, GPT-5, being developed in the labs at a company called OpenAI. Its increased potential has scared a few developers, with its capability and illusion of self-awareness. So why should we be concerned? One problem is that we may no longer be able to know what is real. Then, as we become more dependent on these AI engines, we may be tempted to take them at face value, without checking their sources or conclusions. This could be fatal when our enemies use them to churn out effortless propaganda that sounds real. Photographs could be faked without any way to tell, except by an expert. 

Worse yet, in the future, AI could become smarter than the human race, actually misleading us, and resulting in catastrophe. For now, it’s very helpful to those who know how to use it, but it also has the potential to cause unemployment in certain sectors, such as art, literature, acting, and programming. It all depends on how it’s trained, and you know that the training won’t include anything considered godly. Matters of faith, eternal life, salvation, and conscience may be marginalized at best, or even ridiculed—with the potential for active hostility on the part of future developers. With one step further, AI robots could become idols or objects of hero worship, when used by ungodly corporations or nations. 

A little history

So how did we get started on this path? AI took off in the 1950s after digital computers were developed during the end of World War II. Mathematicians and computer scientists immediately began thinking about what these machines could do with the proper programming. They quickly came to the conclusion that computer games were the ideal platform for exploring machine intelligence. Then the knowledge gained from experimenting could be used in more practical applications. 

In 1950 Claude Shannon (the father of Information Sciences) of Bell Labs, published a paper titled “How to Program a Computer to Play Chess.” This paper was widely distributed and is still used today by budding programmers. Chess was chosen as a vehicle for experiments, similar to how biologists in training study the fruit fly. Computer chess was ideally suited for the study of machine intelligence for the following reasons: 1) it is an intellectual activity which many people are familiar with, 2) the rules are well defined, 3) there are only three outcomes (win, lose, or draw), 4) the games are repeatable, so parameters can be changed to see what difference they make to the resulting play, and 5) there are grand masters at chess, players of great skill, meaning that the progress achieved by the new programs could be tracked by playing against these “rated” humans. (Bobby Fischer was rated at 2785 in 1972, the highest rating ever given at that time.)

Initially the artificial programs played poorly, at a beginner level (rated 1000). Due to slow hardware, small memories, and immature algorithms, the programs could only look ahead two or three moves in three minutes. However, by 1957 some IBM scientists predicted that due to progress in electronics and higher performing algorithms, in 10 years’ time chess programs would actually be able to beat grand masters (rated 2400 or above). They were seriously mistaken about the difficulty of this task, and some grand masters were predicting that a computer program could never beat a world champion. Nevertheless, IBM saw a marketing opportunity: If they could produce a machine that could beat the world chess champion, their world-wide profile would be enlarged tremendously. So they hired a team of graduate students from Carnegie Mellon University to develop a chess program they dubbed “Deep Thought” and run it on an IBM supercomputer. IBM later renamed it “Deep Blue.” It proved to be a block-buster.

The new program consisted of 512 processors with special electronic circuits for generating legal chess moves; it could also evaluate the resulting positions and choose the best ones. They hired a grand master to critique the program and tune its evaluation parameters. Deep Blue could look at least 12 moves ahead in three minutes of “think time” and evaluate 200 million chess positions per second. Surely this would be triumphant, they thought, believing that no human could withstand such brute force computation.

So in May of 1997 Deep Blue played a match of six games against then world champion Garry Kasparov, rated 2845. The computer won two games, lost one, and drew three games. Kasparov was deeply humiliated; he refused to believe that a machine could play this well and accused IBM of cheating by using humans behind the scenes to recommend moves. Consequently, IBM dismantled the machine and disbanded the team. (Years later Kasparov apologized for his comments, in his book Deep Thinking.)

An AI developer and writer, Jonathan Schaeffer, summed up the match in this way: “Of course, a loss by Kasparov is not a defeat for mankind, but a triumph: man has found a way to master complex technology to create the illusion of intelligence. The press is the gullible audience that has been fooled by a sleight of hand.” 

In Part II we will bring the story up to date with more recent developments and look at some concerns for Christian believers. ■  

Retired from a career in mainframe computer performance and capacity planning, Ken and his wife Marea are parents of three adult children and enjoy their four grandchildren, all of whom are living in Nova Scotia. 

Latest Posts