Key takeaways
- Former CEO of Google, Eric Schmidt, warned about the security vulnerabilities surrounding AI models this week.
- Schmidt did also praise the technology and its potential, saying that it is currently “underhyped.”
- “While AI does present security risks for businesses, there are measures that companies can put in place in order to prevent attacks.”
Former CEO of Google, Eric Schmidt, has expressed concern over the consequences of AI systems being hacked, during an appearance at the Sifted Summit this week.
On the other hand, Schmidt also praised the technology, and said it wasn’t getting the hype it deserved. This comes as many businesses continue to implement the technology.
While AI does present new security risks to businesses, there are ways that companies can keep an eye on their systems and prevent an attack.
Ex-Google CEO Eric Schmidt Warns About AI Hacking Dangers
Google’s former CEO Eric Schmidt has expressed concern over the hacking vulnerabilities of AI technology, during a fireside chat at this year’s Sifted Summit. When asked whether AI is more destructive than nuclear weapons, he warned against “the bad stuff that AI can do.”
In particular, Schmidt warned about the potential dangers of AI falling into the hands of bad actors and being repurposed and reused for something more sinister.
This just in! View
the top business tech deals for 2025 👨💻
Later, he expanded on this idea: “There’s evidence that you can take models, closed or open, and you can hack them to remove their guardrails. So in the course of their training, they learn a lot of things. A bad example would be they learn how to kill someone.”
Schmidt Is Optimistic About AI Future
On the flip side, the ex-CEO also appeared optimistic about the future of AI, and went as far as to call it “underhyped” in its current state. The people who are investing in the technology, he continued, are obviously aware of its potential:
“What I do know is that the people who are investing hard-earned dollars [in AI] believe the economic return over a long period of time is enormous. Why else would they take the risk?” – Eric Schmidt, Former Google CEO
Indeed, many businesses are now focused on implementing AI in the workplace, and AI companies are continuing to develop models to make waves in the enterprise market. However, it isn’t necessarily the case every time that businesses are implementing the tech because they understand its future economic return.
In fact, one study found that some leaders are spending on AI because of FOMO, rather than identifying areas where the technology could directly benefit them. And, while some studies have shown that AI is having some economic impact, others have found that it isn’t having any at all.
Should Businesses Be Worried About AI Systems Being Hacked?
There is plenty of evidence to show that AI models are susceptible to attacks, including jailbreaking attacks, where the AI’s responses are manipulated by a bad actor, so it ignores its safety regulations and produces potentially dangerous content.
However, there are many precautions modern-day businesses can follow in order to keep their model safe. This includes writing a clear and detailed AI plan before deployment, testing and reviewing your model often (and against various potential attacks), and limiting the distribution of sensitive data.
To adopt AI safely, you can check out our dedicated guide here.
