135. Mic Drop: Oren Etzioni has a way to help us sort fact from AI fiction

From recorded future news and prx, this is click here. Were you a stand up comedian?

Someone told me that I am not a professional stand up comedian, but I do try to be an amusing comedy computer scientist.

So maybe he's setting the bar pretty low. From Recorded Future News, I'm Dena Temple Greston and this is Click Here's Mic Drop, an extended cut of an interview we think you might want to hear more of. A few weeks ago, we reported on Russia's Doppelganger. This was a loose confederation of Russian intel agents and contract disinformation operators who have not just been trying to roil elections around the world, but have also been taking aim at institutions of democracy too. And one of the new weapons in their arsenal, super convincing AI generated videos and audio that can make ordinary people believe something that just isn't true. You've heard of this? They're known as deepfakes. The good news is that AI is really good at spotting them, and we talked about that with Oren Etzioni, a veteran computer scientist who's been working on a tool that he says can tell the difference between a real video and one that AI has dreamed up.

We're using good AI to fight bad AI, and this is something that hasn't been widely available before. And we're bringing it to the fore right now for free.

And it's incredibly easy to use. Stay with us.

From Zuckerman Spater. For five decades, Zuckerman Spater has been a leading litigation firm trusted by in house and outside counsel, businesses and individuals. With offices in Washington, Baltimore, New York and Tampa, Zuckerman Spater serves clients throughout the country and around the world in high stakes disputes, enforcement matters and investigations. And at trial, when the lawyer you choose matters Most. Online@Zuckerman.com.

I'm Anna Lynn Guzik, founder of the Anti Nihilist foundation, publisher of the Conversationalist. For years, the Conversationalist has been a home for stories about reproductive rights, climate change, LGBTQ rights, labor identity, friendship, motherhood, and foreign policy. Stories that don't get told enough or told honestly enough in a media landscape dominated by corporate interests. Now we're bringing those stories and voices directly to you with season two of the Conversationalist podcast. We're truly independent nonprofit media. No overlords, no corporate agenda. Just stories that matter. This is journalism as conversation. This is storytelling as resistance. This is the Conversationalist Podcast. Subscribe now wherever you get your podcasts.

I'm Dina Templewrest and this is Click. Here's Mic drop. So if you were talking to your younger self 20 years ago, what were you telling yourself about AI then?

It used to be very difficult to get it to work, to get it to understand what you're saying, to get it to understand visual images. But I think what I would tell my younger self is have faith in the exponential growth of computers. Computers have gotten a million times more powerful, and AI for the first time is remarkably powerful.

In other words, AI has gotten really good. Warren Nione has been working on AI for decades, and despite the challenges in early years, he's always been one of those AI optimists.

AI technology has absolutely outstripped our expectations. All that has changed very quickly in the last few years with generative AI.

Generative AI, these things called large language models, which basically gobble up tons and tons of information and then draw patterns from it. For AI optimists like Oren, the possibilities seemed endless.

When you look broadly, you see just these huge opportunities to save human lives. We see some of humanity's thorniest problems, like climate change. And I see the potential of AI to help there and to help make science proceed better and faster.

But AI has emerged with a darker side, too.

All that said, we have become increasingly aware of substantial challenges with privacy, with jobs, with deepfakes.

Oren helped establish something called the Allen Institute for AI. It was a nonprofit funded by one of Microsoft's founders, Paul Allen. And he was one of the first researchers to warn that AI could have a darker side and could do things like accelerate the spread of disinformation online.

About a year ago, I stepped down and took a bit of a sabbatical and asked myself, what is the most important AI problem that I can be working on today? And I decided in 2024 to dedicate myself to launching AI deep fake detection Tool.

The tool can detect doctored images, audio and video. It's available on truemedia.org, the nonprofit Orin founded to expose deep fakes and politics. And it's really simple to use.

Upload a social media post from TikTok, YouTube, Facebook, X, ET cetera, et cetera. And we analyze the media in it, either audio or video or an image, to determine whether it's been synthesized by AI so it's fake, or it's been manipulated by AI in various ways. And already well before the fall, where the election really heats up. We've seen many cases where people have been using AI to mislead voters in.

The election, like that Joe Biden robocall recording in January that told voters not to bother going to the polls, quote a bunch of Milwaukee. Another had Donald Trump supporting Imran Khan, the former prime minister of Pakistan. I will try my best to get.

Imran Khan out of jail.

We are very worried that these kind of deepfakes, as they're called, are a sign of what's to come, and we are hard at work getting ready to fight that.

So far, 41 states have banned electoral deepfakes and aid states have enacted laws to regulate them. And according to the Washington Post, the U.S. senate is pursuing legislation that would ban deceptive AI from being used in political ads, which is a good start, Oren says. But laws and regulations like these are pretty toothless. If there isn't an easy way to tell quickly what's real and what's not.

There's a lot of conversations about regulatory reform, particularly at the state level. The FCC and the FTC have been passing rules. It turns out that if you want to have rules about deepfakes, AI based deepfakes, you also need the tools to detect whether you're encountering a deepfake or not.

Which is why Oren made the True media tool free.

TrueMedia.org is very much counting on a grassroots effort. We are promoting our free tool to journalists, to fact checkers, to people who have a pulpit that allows them to reach the general public, warren said.

We shouldn't count on social media companies to do this for the rest of us because they just don't see the upside of doing so.

They are promoting engagement. They like things that go viral that cause people to be outraged and to respond. And so the social media companies overall have the profit motive, not the social responsibility motive.

So this problem of being duped by generative AI is not going away. Our laser focus through November is on the elections. Not just ours, but elections in Europe and in India too.

And what I like to emphasize is that AI is a tool and the choice of how it's used is ours. And by ours I mean our society. It's not the choice of Mark Zuckerberg. It's not the choice of the technology itself. AI is not a creature with hopes and wishes and aspirations. It's a tool, and it should not be the choice of computer scientists. We as a society need to have the dialogue and the democratic process to decide what impact AI will have on our lives.

From recorded future news, this has been Mike Drop. It was written and produced by Lucas Riley, Sean Powers, Kat Schuchnak and me, Dina Templerest. We'll be back on Tuesday with an all new episode of Click Here. Have a great weekend.

This show is supported by Odoo. When you buy business software from lots of vendors, the costs add up and it gets complicated and confusing. Odoo solves this. It's a single company that sells a suite of enterprise apps that handles everything from accounting to inventory to sales. Odoo is all connected on a single platform in a simple and affordable way. You can save money without missing out on the features you need. Check out odoo-o.com that's o d o.

O.Com looking for more of the cybersecurity and intelligence coverage you get on Click Here? Then check out our sister publication, the Record from Recorded Future News. You'll get breaking cyber news from reporters in New York, Washington, London, and Kyiv, among others, and you'll see for yourself why it attracts hundreds of thousands of page views every month. Just go to the Record Media.