Image credit: CHLOE ELLINGSON/REDUX
Does the name Geoffrey Hinton mean anything to you? It might not; he never won an election, a Grammy, or a Super Bowl ring. But in the world of cybertechnology and artificial intelligence (AI), he’s more than a rock star, and he made a lot of headlines in May 2023 when he left his job at Google and began publicly expressing his alarm about recent developments in AI. Will Knight, who covers AI as a senior writer for Wired, spoke with Hinton and wrote this May 2023 report.
Read it here.
EXPLORE, REFLECT, SPEAK UP.
1. Hinton says he had two recent flashes of insight that alarmed him about the velocity of AI development. What were they? Why were these two insights so alarming? Explain briefly.
2. In his report on a conversation with Geoffrey Hinton, Knight begins by describing a video clip of Snoop Dogg during a panel discussion. Why might Knight have chosen that opening? Do you think it’s effective? Does it matter that Hinton himself sent the video clip to Knight? Why or why not? (To address these questions, try the ‘what if’ technique: What if Knight hadn’t mentioned that video? Would I have read the article in a different way? What if Knight didn’t mention that Hinton turned him on to the Snoop video? Would I have a different opinion of Hinton?)
3. LET’S TALK. Are you finding all the technical information in Knight’s article a little daunting and hard to comprehend? You’re not alone. Do a little “crowdsourcing” with your classmates. Identify which of you have a pretty fair grasp of the technical information and which of you are more foggy. Work together to expand everyone’s level of understanding. Everyone in the groups should have some research responsibility; don’t just rely on those who know a lot to provide all the answers. And remember, sometimes the most interesting questions come from those who start out knowing the least.
4. AND NOW WRITE. First, think of a short joke, maybe one that you’d tell your nine year-old cousin. Next, do what Hinton did: ask ChatGPT to explain why the joke is funny. Finally, write an essay evaluating and critiquing ChatGPT’s explanation. How accurately did ChatGPT interpret the joke? How reasonable was its explanation? Did you need to ask any follow-up questions? In your report, include the full ChatGPT prompt and response.
LetsTalkLibrary 