We have updated our Privacy Policy and Terms of Use for Eurasia Group and its affiliates, including GZERO Media, to clarify the types of data we collect, how we collect it, how we use data and with whom we share data. By using our website you consent to our Terms and Conditions and Privacy Policy, including the transfer of your personal data to the United States from your country of residence, and our use of cookies described in our Cookie Policy.
{{ subpage.title }}
Is AI's "intelligence" an illusion?
Is ChatGPT all it’s cracked up to be? Will truth survive the evolution of artificial intelligence?
On GZERO World with Ian Bremmer, cognitive scientist and AI researcher Gary Marcus breaks down the recent advances––and inherent risks––of generative AI.
AI-powered, large language model tools like the text-to-text generator ChatGPT or the text-to-image generator Midjourney can do magical things like write a college term paper in Klingon or instantly create nine images of a slice of bread ascending to heaven.
But there’s still a lot they can’t do: namely, they have a pretty hard time with the concept of truth, often presenting inaccurate or plainly false information as facts. As generative AI becomes more widespread, it will undoubtedly change the way we live, in both good ways and bad.
“Large language models are actually special in their unreliability,” Marcus says on GZERO World, “They're arguably the most versatile AI technique that's ever been developed, but they're also the least reliable AI technique that's ever gone mainstream.”
Marcus sits down with Ian Bremmer to talk about the underlying technology behind generative AI, how it differs from the “good old-fashioned AI” of previous generations, and what effective, global AI regulation might look like.
Watch GZERO World with Ian Bremmer every week at gzeromedia.com/gzeroworld or on US public television. Check local listings.
- ChatGPT and the 2024 US election ›
- Can we trust AI to tell the truth? ›
- Ian interviews Scott Galloway: the ChatGPT revolution & tech peril ›
- Emotional AI: More harm than good? ›
- Podcast: Getting to know generative AI with Gary Marcus ›
- Artificial intelligence: How soon will we see meaningful progress? - GZERO Media ›
- Will consumers ever trust AI? Regulations and guardrails are key - GZERO Media ›
- UK AI Safety Summit brings government leaders and AI experts together - GZERO Media ›
- Top stories of 2023: GZERO World with Ian Bremmer - GZERO Media ›
Elon Musk wants a way out of Twitter
Ian Bremmer's Quick Take: Hi everybody. Ian Bremmer here. A Merry Christmas to you all. A happy Hanukkah, just kicked off. Happy holidays to everybody. I'm delighted to close out the year with a Quick Take, getting us kicked off this rather cold, blustery, very bright sunny day in New York. Hence the sweater, it feels like a layered kind of day. And with everyone talking about the meltdown that is occurring every day on Twitter, I might as well weigh in.
Most recently, Elon Musk, the owner, the CEO, not the founder of Twitter, asking everyone online should I step down as CEO saying, "I will abide by the results of this poll." The answer 57.5% saying yes, 42.5% say no. They want him to step down. Probably a lot of Tesla shareholders weighing in on that. Let's not pretend that this is in any way a real or useful poll. You can of course, vote all you want with your burner and your fake accounts. What happens if the 12 hours of the poll happen to be 12 hours when you are mostly sleeping, depending on what your time zone is around the world? Well, you are kind of out of luck. I mean, you snooze, you lose. That's what they say. Not to mention the bot problem, and all of the people on Twitter that aren't really Twitter, they aren't really people. Of course, they get to vote too. It's all performative. Of course, Rasmussen had Elon ahead by four, and they turned out to be a little bit wrong, but that happens frequently.
No. Look, clearly, he wants a way out, otherwise, he wouldn't be posting something like this. And it seems like a good idea. I mean, my God, this guy is the CEO of three different major corporations. I'm the founder of one that's considerably smaller, and I'm not CEO because I understand that if you actually want to focus on the things that you're good at, in my case understanding global affairs, you can't also run a company at the same time. Now, look, maybe Elon has extraordinary energy and focus, and Lord knows he's an incredible entrepreneur, but running three separate companies, one of which you didn't found and you don't understand the inside of, as opposed to the other two, where he really does understand the nuts and bolts and the substance of it incredibly well, is clearly a bridge way, way, way too far. He's just surrounded by too many people that won't tell him no or are incapable of telling him no or just blowing sunshine up his ass. And that means you're much more likely to make bad decisions and deciding to run Twitter as clearly one of them.
Tesla stock is down some 60% from the highs. I saw it bounced, I think it was 5% when the Twitter poll results came in, because as I said in the opening, those are people that really want him back focusing on a world-changing corporation, which is kind of a big deal. I do think, by the way, that there's a very, very serious free speech problem out there, including on Twitter. But I don't think that's an Elon problem. I think that's a business model problem. That's a social media problem. That is a problem that comes from privately owned platforms, where the people that participate are the product. The only way you make money is by maximizing the time that is being spent by the product on the platform. So you can sell that product and sell that data to advertisers and other corporations.
That is not a model that in any way would promote free speech. In fact, in many ways, it's antithetical to free speech. The A/B testing that needs to happen on those platforms is about how do you addict the product? How do you ensure that the product spends more time on those platforms? Also, how do you bolster those platforms, even if it's not with people? So you're not incented to verify your accounts, you're not incented to take bots out because that artificially inflates your numbers and improves the output, improves the business model. None of which is good for free speech.
Now, I do think that Twitter's gotten a little worse since Elon has taken over, in part because check marks no longer really mean anything. It used to be that a checkmark was a sign that whoever it was that was in charge of moderation in Twitter believed that you were a public figure of note, meaning a celebrity or an expert or a corporation or whatnot, but it had some level of importance. It was a quick pneumonic that you could use to say, oh, I should pay more attention to this. Now, today, having a check mark either means that, which is again far from an ideal way of determining who's important or not, but it at least is an explainer. It can also mean, no, you just paid $8 and you could be absolutely anybody to have your Twitter blue. The only way you figure that out is if you go and click a second time to see which it actually reflects. Of course, that is something that most people aren't going to do and even if you do do it, it's after you've already gotten that internal sense that, oh, I should pay more attention. So that makes this less efficient as a platform.
Also, a lot more ads are showing up and you don't know it's an ad until at the bottom when you see promoted by. So again, the first experience you have is, oh, this is content I want to see and then you find out it's not really content you want to see. So I think at the margins, it's gotten a little bit less product friendly over the last four or five weeks, but the free speech experience is largely the same from my perspective. I will say it was kind of cool that Taylor Lorenz was suspended for 12 hours since people got enormously excited about that. Not at least of which was Taylor Lorenz. But okay, that's, that's a cheap thrill for those of you that are on social media. It's not seriously exciting or meaningful.
Nor am I expecting government intervention to matter. I think it's entirely too difficult. It's possible that you'll see TikTok have significant regulations against it, though I doubt it'll be removed. But that has nothing to do with social media problems for free speech that just has to do with the fact that it's Chinese-owned. And a lot of people are concerned about the Chinese government having access to that data.
No, the big question for me is the future of decentralization. Right now we have a governance model where you have a small number of corporations and the individuals that own them that have access to all of this data, might personalized artificial intelligence in the near term, undermine that model when you have individual bots that work for you that can scour the entire internet and give you a news experience, a media experience, a social media experience that you curate as opposed to the sovereigns in the tech space curate that. That would be a radical transformation of the space that would deeply threaten those that own these social media and other tech platform monopolies.
The other broader, longer-term question would be whether the blockchain undermines data surveillance by taking privacy and giving it back to individuals. I think the former might be three years away. The latter is probably more like 10 or 20, but at least for the foreseeable future, this is a very serious problem.
So that's a little bit for me, my take on everything, Twitter, social media, and democracy. And I'll talk to y'all real soon.
Will Elon Musk have a China problem with Twitter?
Following news of Tesla CEO Elon Musk's plan to buy Twitter, one of the most bizarre and viral reactions came from … his Chinese doppelgänger.
“My man, my man. I love you,” said the self-proclaimed Yi Long Ma, who’s become an internet celebrity for his videos spoofing Musk on Douyin, the Chinese-language version of TikTok. The world’s richest man himself gave “Yi Long” a thumbs-up late last year, joking that perhaps he’s part-Chinese.
His lookalike in China is clearly excited about Musk owning the social media platform. But will the world's richest man’s ties to China hurt him and Twitter? There are two sides to that argument.
On the one hand, China could become a big headache for Musk in his new role. That’s the prevailing Western media narrative, which argues that since Tesla is heavily invested in China, Beijing could influence Musk to relax some Twitter rules that Xi Jinping doesn’t like.
Chinese diplomats and state media outlets have long complained about their accounts being labeled "state-affiliated." Also, Musk's plans to turn Twitter into a bastion of free speech by relaxing content moderation rules and getting rid of bots threaten China's widespread use of fake accounts to push pro-Beijing propaganda and misinformation.
More broadly, Musk, the world's most famous libertarian, will be running a social media platform while still doing business with the world's most powerful authoritarian regime. You don't have to be a jealous fellow billionaire to ask whether there might eventually be some tension there.
What’s more, China has leverage on Tesla. The country accounts for almost a quarter of global sales of the electric vehicle company, which has built a massive data center in Shanghai, China's largest city. If push comes to shove, Beijing has ways to hurt Tesla in the Chinese market — perhaps by regulating in favor of its local competitors, or stealing its data and tech.
"Tesla stock is already under the gun because Musk has sold billions of dollars worth of his shares and is using his remaining stake as collateral to pay for Twitter," says Eurasia Group analyst Scott Bade. "China is important to Tesla, not only as a market but also as a source of rare earths for its EVs, and there are fears it could use both to influence Musk on Twitter."
Interestingly, when Shanghai entered its COVID lockdown over a month ago, we didn’t hear a peep from the billionaire, who earlier cited California's pandemic restrictions — and high taxes — to justify moving Tesla’s US headquarters to Austin, Texas.
On the other hand, maybe these China jitters for Musk are overblown. Why? For one thing, Twitter, which has been blocked by China since 2009, had barely 10 million daily active Chinese users (via VPN) in 2016, the last time the US-based platform checked. Compare that to the whopping 573 million on Sina Weibo, China’s answer to Twitter — which Chinese censors fully control.
For another, any gesture perceived as kowtowing to China would pose huge risks for Tesla and Musk himself in the US, where Twitter is most influential. In today's highly polarized US political environment, being anti-China is perhaps the only thing the extremes of Twitter can agree on. Musk certainly loves a good culture war, but getting caught on the wrong side of the US-China rivalry could lead many Americans to stop buying Teslas.
Also, Twitter, although growing in importance, is still hardly a big part of China’s broader social media strategy to win over foreign netizens.
"Musk is unlikely to be under major pressure to quash 'anti-China' voices on Twitter, or grab the personal data of Chinese dissidents in the near term," says Bade. "Beijing has many other tools available to counter negative news and track its enemies."
Finally, China needs Tesla as much as Tesla needs China, at least for now. Tesla has invested billions of dollars there, and its presence has spurred local EV makers to become more competitive. China wants to become an EV powerhouse by riding on Tesla's coattails — and it’s hard to imagine Xi will let Twitter get in the way of that as the Chinese economy keeps sputtering due to his zero-COVID policy.
What do you think? Let us know here.
Personal data risks with TikTok; Tesla driverless cars investigation
Marietje Schaake, International Policy Director at Stanford's Cyber Policy Center, Eurasia Group senior advisor and former MEP, discusses trends in big tech, privacy protection and cyberspace:
Beijing took a stake and a board seat in TikTok owner ByteDance's key Chinese entity. Should I worry about my data on TikTok?
Now, being concerned about where your data ends up is always a good idea, but for underage children, many of whom love video-sharing apps and social media, that question is even more sensitive. And for apps that end up being accessible by governments, and essentially most of them are, you want to be aware of what you share. I recall an account of an American teenager being shut down as they highlighted the human rights violations of Uyghurs in Xinjiang, which is, of course, something that should be highlighted and it's troubling that the video-sharing company intervenes on behalf of a Chinese state agenda.
Tesla is under formal US investigation into Autopilot crashes. Will driverless cars also be more vulnerable to cyberattacks?
Now the simple math suggests yes, because the more software, the more risk. Any smart device, whether it's a fridge, a phone, or a car, can be vulnerable to hacking and attacking. But with self-driving cars, with a lot of software, the risk of attacking or commandeering is easy to imagine. Slamming a hacked car into a crowd can be the dream of terrorists. So, getting more transparency about risks, liability and responsibility for software quality, the security as well as updates are all much needed if trust in the promise of self-driving cars is to lead us to a positive place.
Sri Lanka Blocks Social Media: Tech in 60 Seconds
Should Sri Lanka have blocked social media following the terror attacks?
That's a hard one. Misinformation spreads on social media and there's an instinct to say, "Wait, stop it!" But a lot of useful information also spreads and people get in touch with each other. So I would say no they should not have blocked it.
Are Tesla cars at risk of exploding?
There was one video from China of a parked Tesla exploding. I don't think you really have to worry about it though. I am curious to know what that video was really about.
Why do tech companies hate the census citizenship question?
Because if you ask people whether they're citizens. A lot of people will answer and you'll get bad data and the card companies need to know where they set up their operations. Good data matter to Silicon Valley.
What happened during the Space X Crew Dragon accident?
We don't know this one for sure either but one of the engines in a SpaceX test exploded. No one was hurt. Let's hope it was something to do with the way it was set up - not something deep and systematic.
And go deeper on topics like cybersecurity and artificial intelligence at Microsoft Today in Technology.
How to Spend Your Performance Bonus: Money in 60 Seconds
Here are some important tips to use your bonus wisely.
It's Money in 60 Seconds with Sallie Krawcheck!
And go deeper on topics like cybersecurity and artificial intelligence at Microsoft on The Issues.
Tech Policy in 2019, Facebook, Tesla and AI News
The midterms are over and that means less irrational tech policy in 2019. But it could also mean less rational tech policy too.It's Tech in 60 Seconds with Nicholas Thompson!Take action against cyberattacks and add your name to the Digital Peace Now petition: https://digitalpeace.microsoft.com/#gf_1