Bill Gates shocked as ChatGPT aces AP Bio exam

There was a time when Bill Gates had doubts about artificial intelligence technologies like ChatGPT. But, Gates was “shocked” when ChatGPT recently passed an AP Bio exam. 1000

Even Bill Gates was shocked by the quick development of artificial intelligence technologies like ChatGPT.

On Thursday’s session of his “Unconfuse Me” podcast with Khan Academy CEO Sal Khan, Gates described “the most stunning demo I’ve ever seen in my life”—ChatGPT’s performance on an AP Biology test in August 2022.

Gates tested OpenAI’s AI-powered chatbot in June and was underwhelmed. “I said, ‘Yeah, it’s kind of an idiot savant. Gates stated, “I don’t think it’s practical.”

He gave OpenAI CEO Sam Altman a challenging challenge: bring ChatGPT back to him once it could demonstrate advanced, human-level capability by scoring the highest on the AP Biology test.

“I thought, ‘OK, that’ll give me three years to work on HIV and malaria,’” Gates joked on the program.

Gates witnessed ChatGPT earn a perfect five two months later when OpenAI’s developers returned. Gates said the software could read and write “in the sense humans do” now.

“Wow, it is so good,” Gates exclaimed.

Gates noted that even the most capable AI models may make big mistakes or fake facts.

“Let’s see where we can put it to good use,” Gates replied, pleased by the quick progress. AI might “close the gap” in education.

Gates and Khan debated education’s “good use” on the podcast.

Gates, an optimist, believes AI chatbots like ChatGPT or Google’s Bard might start teaching youngsters to read and write next year. Khan said his business is creating Khanmigo, a chatbot trainer, for that reason.

Khan stated Khanmigo may “act like a fairly good human tutor”. “It has moments that would pass the Turing test, where you’d think there’s a good human on the other side of the chat.”

However, brilliance doesn’t guarantee human-level performance. Khan and Gates said AI tutors won’t replace instructors, but they may make teachers’ work simpler and offer cheaper tutoring to underprivileged pupils in low-income communities in the U.S. and abroad.

Khan stated GPT-4, OpenAI’s newest big language model, makes blunders. However, OpenAI claims the model is “dramatically better” than prior versions and can outperform 90% of human SAT test takers.

Last month, Barnard College child psychologist Tovah Klein warned CNBC Make It about those blunders. She suggested AI models should explain their responses better and not become children’s exclusive source of information.

“I think AI has a role if we have the science showing that that kind of learning, in addition to a teacher, is useful,” Klein said. “We don’t really know.”

Gates said he has seen enough of ChatGPT’s progress, particularly its rising capacity to explain its thinking, to foresee a game-changing influence on education within a decade.

Gates said, “If we think about the next 10 years, [in terms of] both the absolute level of learning and the gap with lower-income, minority students… these new tools can both close the gap and raise the overall level of achievement.”

Conclusion:-

Bill Gates was shocked by the rapid development of artificial intelligence technologies like ChatGPT. During a podcast with Khan Academy CEO Sal Khan, Gates described ChatGPT’s performance on an AP Biology test in August 2022. Initially skeptical, Gates challenged OpenAI CEO Sam Altman to bring ChatGPT back to him once it demonstrated advanced, human-level capability. After two months, ChatGPT scored a perfect five, reading and writing “in the sense humans do” now. Gates believes AI chatbots like ChatGPT or Google’s Bard might teach youngsters to read and write next year. Khan Academy is creating Khanmigo, a chatbot trainer, for this purpose. However, brilliance doesn’t guarantee human-level performance. AI tutors won’t replace instructors but may make teachers’ work simpler and offer cheaper tutoring to underprivileged pupils in low-income communities. OpenAI’s GPT-4, its newest big language model, makes blunders, but OpenAI claims it outperforms 90% of human SAT test takers. Child psychologist Tovah Klein warned CNBC Make It about AI models’ blunders, suggesting they should explain their responses better. Gates believes ChatGPT’s progress could have a game-changing influence on education within a decade, closing the gap and raising overall achievement.