UC Berkeley’s Hany Farid: “With AI, we’re making the same mistakes that we did with social media”

The Times Tech Podcast

Will Morley

Business, Unknown, Technology

4.9 • 654 Ratings

🗓️ 9 June 2023

⏱️ 40 minutes

🔗️ Recording | iTunes | RSS

🧾️ Download transcript

Summary

The Sunday Times’ tech correspondent brings on Hany Farid, of UC Berkeley’s School of Information, to talk about the new age of AI (artificial intelligence) misinformation (4:00), the proliferation of new tools (6:40), AI doomerism (9:50), regulation (13:10), history repeating itself (15:40), the AI election (18:20), simple ways to regulate AI (22:45), why imposing rules is easier than the industry would have us believe (28:50), the challenge presented by the open-source approach (30:30), and how the 2024 election could go wrong (34:00).

Hosted on Acast. See acast.com/privacy for more information.

Transcript

Click on a timestamp to play from that location

0:00.0	Yo! Technology!
0:02.0	What is it all about?
0:04.0	Hello and welcome to Danny in the Valley.
0:07.0	Your weekly dispatch from behind the scenes and inside the machines that can now very easily clone my voice.
0:13.0	In fact, what you are listening to right now is an AI-generated clone.
0:18.0	I'm not kidding.
0:20.0	... I'm not kidding. Hello, hello, hello, I won't be doing my normal intro today because my Danny bot just did it.
0:42.3	That's wild, right? So that was actually a bot that you just heard.
0:46.7	Wasn't quite me, but it was pretty close. And I want to get to how that happened in just one second.
0:52.7	But I want to first introduce who we have on
0:55.4	the program today, who is one of my favorite people to talk to about the internet, misinformation,
1:00.9	social media, and generally a landscape of online life. And that is Hani Farid. He is a professor
1:06.1	at UC Berkeley who spends his waking hours analyzing misinformation, the latest tools that are making
1:11.7	it always easier to develop.
1:14.0	He also spends a lot of time with lawmakers on Capitol Hill about how to rein in the
1:18.1	worst aspects of all of this stuff.
1:19.5	And as you will hear, when we're talking, he mentions in our conversation just how
1:24.1	easy it is to clone people's voices now.
1:26.7	So in the context of politics, the context of
1:29.5	fraud, this is pretty unsettling. So after we got off, I tried it. And so there's a startup
1:34.7	called 11 Labs. And for five bucks a month, you just upload any voice recording that all has to be
1:42.5	at least one minute. You type in whatever you want that
	...

Please login to see the full transcript.

Previous episode | Next episode

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

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