I know you’ve probably heard enough about ChatGPT for the next few years, but allow me to share my thoughts while you still have some guarantee it’s a human writing this blog. There's more to discuss than ChatGPT, of course, a popular topic being the apocalyptic AI revolution. But let’s start with the newest writing program...
I took a test a couple months ago to see if I could distinguish between writing assignments composed by ChatGPT and those by actual students. For a few, I honestly could not tell the difference. The program knew exactly how a typical eight-year-old student would write, mistakes and all, making it incredibly difficult for your typical teacher who’s marking assignments (many teachers who took the test couldn’t tell the difference).
Google Translate was already an issue for the French immersion program at my daughter’s school, where assignments composed in English can be translated into near-perfect French. Now there's a program to write the original essay! High schools and universities are grappling with what to do with assignments that computers can now effectively write in seconds.
Then there's robotics... Only a few years ago, some of us were casting doubt on the future of driverless cars, but they’re a reality in cities like Phoenix and San Fransisco. One company runs 10,000 driverless taxi trips a week. Sure, there are glitches – like cars suddenly stopping in the middle of the road on a foggy day – but give it a few years for developers to solve these issues. Our 2017 Honda already has safety features to keep me on the road when I take my hands off the steering wheel (mostly for stretch breaks). What’s not a stretch is to foresee computer drivers becoming safer than humans.
Even more amazing: Google is experimenting with robots who are teaching themselves how to play soccer. The videos are worth watching: https://ca.finance.yahoo.com/news/ai-robots-figure-play-football-132948589.html. In this innocent experiment, one can laugh at the robots falling and picking themselves up as they compete against one another – appearing much like five-year-olds learning the game. In other contexts, such as modern warfare, one can envision some frightening scenarios.
Like most technologies, artificial intelligence offers both opportunities and risk. But it's striking how many technologically astute people are sounding the alarm bells. By their assessment, AI could soon pose the same risk to humanity as nuclear weapons.
It took me a while to fully grasp this concern. What exactly are we to fear? I mean, I’ve seen so many movies with machines destroying humanity, I’m not sure how seriously to take it anymore. (Full disclosure: I recently watched Terminator 2 and yes, it was still epic.)
Let’s be clear, there are a lot of unknowns. But the basic fear is that a super intelligent computer could very well become smarter than humans – most believe it’s simply a matter of time. At that point – maybe fifty years from now – it becomes difficult to know or even trust whether the AI will serve humans’ or its best interests. Or whether it might misinterpret or mishandle a positive human directive; to protect the future of the earth, for example, it might just destroy all humanity (a little extreme, but you get the point).
It might help to think of an adult talking to a three-year-old. A three-year-old's level of understanding/intelligence is such that an adult could easily trick him/her into doing something that’s not in the child’s self-interest. Most adults wouldn’t act in this way, but we know some will. Once a machine develops an intelligence where it understands human nature even better than we do, it could become the adult in the room while we remain the child. One could foresee acts of deception that may not be in our best interest, for as we know, humans are susceptible to belief in all kinds of disinformation.
Combine this with perpetual self-learning and I start to understand why some technologists are getting antsy. Just think about those robots teaching themselves how to play soccer. Imagine that ability times 100, when it becomes the norm for robots to teach other robots any task – including making other robots. I don't think it will end with soccer.
There are endless possibilities for which one’s mind can imagine (again, I just watched Terminator 2, so I can imagine). Think about how limited human intelligence has transformed the world in a relatively short time period. What could super intelligent beings do?
All that said, a doomsday scenario is not inevitable. As a wise boy once said in a movie I recently watched: “The future’s not set. There’s no fate but what we make for ourselves.”
One can hope we will learn quickly and harness this new technology for the good.
No comments:
Post a Comment