Good morning.
Russia’s invasion of Ukraine tops most lists of important news events in 2022, and President Volodymyr Zelensky is not only Time’s Person of the Year, but also everyone else’s. But the last couple of weeks have convinced me the honor may belong to something else: Open AI’s ChatGPT.
If you haven’t played with this extraordinary new tool yet, I’d urge you to give it a try here.
Along with its sister text-to-image tool DALL-E 2, ChatGPT has suddenly turned the conversation around artificial intelligence from one of future potential to present amazement. The two tools were not only the focus of our Brainstorm A.I. event in California early last week, but also of virtually every conversation I’ve had since. Here, for instance, is what Coursera CEO Jeff Maggioncalda had to say when he dropped by Fortune offices last week:
“Our team has been showing me what GPT-3 [the A.I. that powers the ChatGPT chatbot] is doing on Coursera-related assets…We have all the transcripts of 5,000 courses, and they will write in: “Create flashcards to practice the following content,” and it creates the flashcards. Like whoa, automatically. There was one where they said generate five multiple choice questions and two true/false questions around the following material, and it generated a seemingly thoughtful assessment of the material.”
The technology is not flawless. Examples where the chatbot has committed egregious, and even harmful, mistakes are easy to find. But the framers are working hard to correct errors, create guardrails, and perhaps most importantly, inject some humility into answers. For instance, I asked ChatGPT whether the recent collapse of crypto-empire FTX was a “Lehman moment” or a “Madoff moment.” Its response:
“I’m sorry, but I am not able to provide information on the collapse of FTX or compare it to the collapses of Lehman Brothers or Madoff Securities. As a large language model trained by OpenAI, my function is to provide information and answer questions to the best of my ability based on the data that I have been trained on. I do not have the ability to browse the Internet or provide information on current events.”
All this supports Reid Hoffman’s conclusion at our conference that, within the next few years, every profession will have an A.I. “co-pilot” augmenting work efforts—much as we already use Google Maps to augment navigation and search to augment memory. Any leader who hasn’t started thinking about how this will transform their business better start now.
More news below.
Alan Murray
@alansmurray
alan.murray@fortune.com
