OpenAI Brings Scarily Convincing New Chatbot to the Public

KQED's Forum

KQED

News, Politics, News Commentary

4.2 • 726 Ratings

🗓️ 13 December 2022

⏱️ 56 minutes

🔗️ Recording | iTunes | RSS

🧾️ Download transcript

Summary

The newest artificial intelligence sensation, OpenAI’s language model ChatGPT, can teach physics, write in verse and even compose a convincing letter from Santa Claus. Trained on billions of written words ranging from online message boards to classic literature, ChatGPT can also hold respectably intelligent – even witty– conversations, all the while refining its accuracy and style with constant user feedback. But critics say the technology can enable cheating and amplify falsehoods. We’ll talk about the potential uses and abuses of ChatGPT. Guests: Nitasha Tiku, tech culture reporter, Washington Post Kevin Roose, technology columnist, New York Times Daniel Herman, high-school teacher and author of the Atlantic article, "The End of High-School English" Learn more about your ad choices. Visit megaphone.fm/adchoices

Transcript

Click on a timestamp to play from that location

0:00.0	Support for KQED Podcasts comes from Landmark College, commemorating 40 years of educating people who learn differently, with programs on campus and online for both students and professionals.
0:12.0	Learn more at landmark.edu.
0:14.7	Switch to Comcast Business Mobile and save hundreds a year in your wireless bill.
0:18.8	Comcast Business, powering possibilities. Restrictions apply. Comcast business powering possibilities.
0:21.4	Restrictions apply.
0:22.1	Comcast business internet required.
0:23.4	Compar's two unlimited intro lines and lowest price 5D plans of top three carriers.
0:26.4	Tax on fees extra, reduce speeds after 30 gigabytes of usage.
0:28.8	Data thresholds may vary.
0:31.1	From KQED in San Francisco, this is Forum. I'm Mina Kim. The latest sensation in artificial intelligence is inspiring both awe and dread.
0:58.0	Chat GPT, a chat bot that can write like a human, built by the California company OpenAI.
1:04.0	Need to craft a personal letter, an essay on Moby Dick, even computer code.
1:08.0	Chat GPT will do it convincingly. It will also have a conversation with you,
1:12.7	complete with self-deprecating humor, and it's free and publicly available online. But what does it
1:18.7	mean to live in a world that lets AI do the talking? This hour we explore the potential uses and abuses
1:24.8	of AI-generated text.
1:37.2	I'm Mina Kim. Welcome to Forum.
1:44.3	When chat GPT opened to the public less than two weeks ago, it was immediately swamped by users, more than a million in the first few days, eager to try the chatbot that can write on seemingly any topic in any style,
1:50.7	a college-level comparative literature essay, physics concepts, and simple terms,
1:55.7	a birthday message to an in-laws sensitive about their age.
1:59.3	The chatbot by San Francisco-based company OpenAI was trained on the vast reserve of
2:04.6	Internet data to sound knowledgeable and witty, human-like.
	...

Please login to see the full transcript.

Previous episode | Next episode

Disclaimer: The podcast and artwork embedded on this page are from KQED, and are the property of its owner and not affiliated with or endorsed by Tapesearch.

Generated transcripts are the property of KQED and are distributed freely under the Fair Use doctrine. Transcripts generated by Tapesearch are not guaranteed to be accurate.