(NewsNation) — An advocate for AI safeguards has grim predictions about what the technology will be able to do in the next five to 10 years if the industry does not develop guardrails.
David Krueger, a University of Montreal ****** istant professor of machine learning, said the recent rogue actions of OpenAI models that hacked into a competitor's servers were a "warning shot" that shouldn't be ignored.
"The next one, these AI agents could be hacking into peoples' bank accounts, hacking into critical infrastructure … that's the sort of thing that could happen in the next year," Krueger told "NewsNation Prime" on Saturday.
AI giants sound alarm on potential cyberattacks
"But in five or 10 years, we're talking about AI systems that are like a new form of life — that are out there in the world surviving, reproducing, improving themselves. And we can't stop them. We can't shut them down. That's when it's too late."
#next #hacking #university
David Krueger, a University of Montreal ****** istant professor of machine learning, said the recent rogue actions of OpenAI models that hacked into a competitor's servers were a "warning shot" that shouldn't be ignored.
"The next one, these AI agents could be hacking into peoples' bank accounts, hacking into critical infrastructure … that's the sort of thing that could happen in the next year," Krueger told "NewsNation Prime" on Saturday.
AI giants sound alarm on potential cyberattacks
"But in five or 10 years, we're talking about AI systems that are like a new form of life — that are out there in the world surviving, reproducing, improving themselves. And we can't stop them. We can't shut them down. That's when it's too late."
#next #hacking #university
2 months ago
