The newsmagazine show “60 Minutes” recently had a series of whizbang interviews with Google and Microsoft executives, examining the pros and cons of artificial intelligence. The conversation was largely philosophical because there is not enough experience with the technology to understand where it might go. The utility of soccer-playing robotic dogs notwithstanding, it was a fascinating hour of television..What struck me most about the discussion was the Monty Pythonesque resort to Orwellian language when trying to understand the penchant for the program to lie in its responses..When asked about inflation, the chatbox BARD, gave a useful response with references to five books. Unfortunately, the books do not exist. To our tech guru masters, this is known as “hallucination”. Try that with the judge the next time you visit him in court. “Sorry your Honor. I wasn’t lying. I was merely hallucinating in my response.”.Just as I cautioned my children that lying leads to distrust, so do I now distrust BARD and all his electronic friends. In this case, the lying was merely funny. But pity law professor Jonathan Turley at George Washington University. A colleague of his had asked ChatGPT for an article about the legal implications of the new AI technology and the article quoted a Washington Post article accusing Professor Turley of sexually harassing one of his students. How is that for irony?.Was it true? Well, no. The Washington Post article did not exist, and Professor Turley had a rock-solid alibi to cover the charge in question. Do I think this technology is benign? Professor Turley no longer thinks so. The owners of ChatGPT have yet to call and apologize for the fabricated story that defamed him. Note to these people… he is a LAW professor. You might want to make that call..A friend who works in the field explains the fascinating technology of large language models as follows,.Computers are calculators, they are not people or 'beings' of any kind. They compute — they do not discern. They are machines that crunch numbers, they do not have insight. They can be programmed to convert numbers to text. We can use them to calculate the statistical associations between text fragments contained within vast accumulations of documents..What underlies so-called "artificial intelligence" (AI) like ChatGPT is nothing more than a large collection of computed probabilities gleaned from word associations and sequences. These probabilities are often based on billions of human-generated textual documents..When you type in a question to ChatGPT or any other language model, what you are doing could be thought of as something akin to pressing the keys on a piano or some other musical instrument..ChatGPT converts your question into a numerical representation and maps that numerical representation against its sea of probabilities — the ones previously derived from a vast universe of documents. Then, it generates a sequence of English text based on where your question fell in its sea of textual possibilities..The collection of probabilities is so large, these AI models can calculate resulting text that is properly formed English and has a coherent and, usually, relevant meaning..Statistical probabilities derived from large corpora of human documents offer valuable insights. But the mystical overtones being used to describe advances in artificial intelligence is starting to get downright weird. I'm afraid much of the shock and awe in the popular press and on social media is beginning to mislead many people into adopting a quasi-spiritual understanding of what is happening in technology..He blames the mysticism that surrounds artificial intelligence on something called “the computational theory of mind” which holds that the human mind is an information processing system and that thinking, and consciousness are forms of computation. Thus, if a machine can compute quickly then it is thinking and therefore it must be conscious. That is Mr. Kurzweil’s singularity argument. It is the spooky moment when machines outthink humans, and we have five seconds to pull the plug on our HAL 9000’s before they destroy us..Sarcasm aside, chat boxes and artificial intelligence are important new technologies that in time, will affect how we live. I don’t think we know how this hallucination thing is going to work out but maybe the solution to the technological threats is built in. If you can’t trust the output and it may cause you to end up in court, reliably accused of defamation, then how often are you going to use the technology? If the time you save in having a response in under five seconds is eaten up by the hours you must spend fact checking every word, then why bother?.If I were king, I would ensure that the owners of the technology are responsible to pay thrice the damages caused by reliance on their technology. For those who want to slow down the introduction of this technology, I think the sphincter-tightening exposure to expensive, court-mandated payouts are just the ticket..Mr. Pichai of Alphabet is talking about installing a “Google” button so that you can search the base documents instead of risking a day in court. Instead, I am going to continue using Wikipedia and check the references myself..Murray Lytle, Ph.D. is a retired engineer living in Alberta. His resource development career took him around the world, and he was one of the last permanent members of the National Energy Board before it was modernized to the Canadian Energy Regulator.
