Trending Articles
Trending Topics
Trending
Sam Altman said what?! The OpenAI CEO has been doing media rounds this week, sparking a series of increasingly alarming headlines based on (perhaps out of context) quotes from the tech leader, such as these:
“I get scared sometimes to use certain AI stuff.”
“There are no adults in the room.”
Oh, and: “I sort of felt like it was the Manhattan Project.”
Here, we’ll reveal what Altman said and what we imagine he actually means, because otherwise he might be bonkers.
On a podcast, Altman admitted he was “scared” using AI some of the time. He reportedly said: “I get scared sometimes to use certain AI stuff, because I don’t know who’s going to have it.”
If only Altman knew someone at an AI company who could answer that question for him. Oh, wait. If the CEO of one of the AI leaders feels this way, imagine how the rest of us feel.
But he was making a good point about how people use ChatGPT as a therapist, life coach or for relationship advice – and that may not prove wise, given AI chatbots don’t have doctor-patient confidentiality. That shouldn’t come as a surprise, given ChatGPT is very much not a doctor. It all means that anything entered into the chatbot could be used against you in court.
And while Altman’s advice to be wary is worth heeding, court cases may well be the real motivation behind what Altman is saying: OpenAI has to keep all data on user conversations, even if you’ve asked it to be deleted. That’s because the information is being collected as part of a lawsuit brought by The New York Times. If you don’t like the data retention, blame those pesky copyright infringement cases.
Then Altman met with Federal Reserve vice chair Michelle Bowman, to tell spooky stories about the potential for AI to spark widespread fraud across the banking industry.
This is largely down to the fact that many banks and financial institutions apparently use voice prints for authentication. It’s like a PIN but for rich people and their money managers.
“I am very nervous about this,” he reportedly said. “A thing that terrifies me is apparently there are still some financial institutions that will accept a voice print as authentication for you to move a lot of money or do something else. You say a challenge phrase and they just do it.”
“AI has fully defeated most of the ways that people authenticate currently, other than passwords,” he added. “Right now, it’s a voice call; soon it’s going be a video or FaceTime that’s indistinguishable from reality.”
Now, there have been examples of fraudsters using AI to imitate voices, so Altman isn’t being totally alarmist.
But it’s also worth noting another project that Altman is involved with: World ID, a biometric identity platform that scans your eye as the basis for a secure digital identity. That’s sort of organised across two of his companies – Tools for Humanity and Worldcoin – with one making an eye-scanning device called the Orb and the other building a crypto-powered system to run it all.
Now, that might just solve the looming AI-sparked fraud fest, eh? Handy that.
Also on that podcast, Altman talked about GPT-5, saying that testing the next AI iteration from OpenAI made him “very nervous”.
He even compared it to the military project that led to atomic weapons: “I sort of felt like it was the Manhattan Project.”
Fraud and data privacy sort of seem rather less important than that.
Of course, this is how AI companies talk about their product: they big it up as world changing or potentially world ending as part of their marketing efforts, and then companies roll it out and employees are left with a nifty coding assistant or largely useless report generator.
Why would Altman build such a technology if he thought it was so potentially dangerous? It’s better to control the tool than let someone else do it. But in the same podcast, Altman said that when it comes to AI regulation, “there are no adults in the room.”
While that certainly rings true, on this point, we’ve got to wonder what Altman means: after all, isn’t he in the metaphorical room?
If the aim is to make us all a bit nervous when asking ChatGPT for recipe suggestions, silly photos and relationship tips, mission accomplished – but if we’ve learned anything from Altman this week, don’t ask the bot for help easing your anxiety.
