Professor Joseph McRae Mellichamp of the University of Alabama, speaking at conference at Yale University to an audience that contained the Novel Price winner Sir John Eccles, famous for his discovery of the synapse, together with a number of the pioneers of AI, said: “It seems to me that lot of needless debate could be avoided if AI researchers would admit that there are fundamental differences between machine intelligence and human intelligence — differences that cannot be overcome by any amount of research.” In other words, to cite the succinct title of Mellichamp’s talks, “‘the artificial’ in artificial intelligence is real.”
What was the last thing you heard about artificial intelligence? Maybe it was about ChatGPT, an open AI web app that invites people to ask the computer to write anything they can think of.
Chris Hutchinson on Twitter asked it to rewrite the Gettysburg Address in the style of the psychedelic funk band Sly and the Family Stone. The AI said it would be disrespectful to rewrite such a historic speech in this style. Then he asked for a rewrite of the speech as a haiku, and the AI complied. Later, another user was able to get the speech in the style of Sly and the Family Stone by wording the request differently (and possibly by his preceding requests). Maybe ChatGPT had a change of heart after refusing the first request.
Educators have been worried that this program (and others produced in its wake) will allow students to task their computers with writing papers for them with minimal chance of detection, but educators are prove to worrying and are probably assuming too much AI language proficiency at this point. Writers worry this program threatens their jobs, and those who work for any of the click-bait sites on pop culture, movies, and games should worry. The garbage prose ChatGPT spits out is totally on par with their daily posts.
You won’t find this in Lennox’s book, 2084: Artificial Intelligence and the Future of Humanity. It was published in 2020. Developments in this field will be fast and fierce (no, frenetic. Wait, it’s fast and feverish, right? Fulminous?) Lennox couldn’t deal with the very latest news, but he does deal with the ideas and claims many in the field of artificial intelligence are making.
Current advances in AI have sparked hopes and fears similar to George Orwell’s 1984, but instead of INGSOC controlling our society, it would be supercomputers that had developed themselves beyond their creators’ imagination. If it came to reality by 2084, supporters ask, wouldn’t it be poetic?
Lennox explains some of the benefits of current machine learning and some of the changes we see coming as robots take over select jobs. For example, the freight industry could be transformed by trucks that drove themselves. (How would they refuel? Could criminals take advantage of them?) He also explains some of the dangers we can already see in AI’s current uses. China’s surveillance state already looks resembles an episode of Black Mirror in which approved behavior and social media influence controls a society of people constantly monitored by unseeing eyes. Facial recognition programs may violate privacy by design and are only as good as they are accurate. False matches have already gotten a few people in trouble.
Continue reading 2084: Artificial Intelligence and the Future of Humanity, by John C. Lennox