OpenAI had touted GPT-5 as having PhD-level abilities, but Google DeepMind doesn’t seem to think too highly of those claims.
Google DeepMind CEO Demis Hassabis has said that claims made by AI companies of having created PhD level intelligences are “nonsense”. “ You often hear some of our competitors talk about these modern systems that we have today are PhD intelligences,” Hassabis said at the All In Summit. “I think that’s nonsense. They’re not not PhD intelligences. They have some capabilities that are PhD-level, but they’re not in general capable, and that’s what exactly what general intelligence should be, of performing across the board at the PhD level,” he added.
OpenAI CEO Sam Altman had earlier claimed that GPT-5 had “Ph.D-level” capabilities. “It’s like talking to an expert, a legitimate PhD level expert in anything,” he’d said at the launch of GPT-5, and called it “the most powerful, the most smart, the fastest, the most reliable and the most robust reasoning model that we’ve shipped to date.”
But Demis Hassabis was quick to point out that current systems aren’t anywhere near PhD-level in general abilities. “In fact, as we all know, interacting with today’s chatbots, if you pose the question in a certain way they can make simple mistakes with even high school Maths and simple counting. That shouldn’t be possible for a true AGI system. So I think that we are maybe, 5 to 10 years away from having an AI system that’s capable of doing those things,” Hassabis said.
Hassabis also said that these systems didn’t yet have the ability to learn like humans. “Another thing that’s missing is continual learning, this ability to — online — teach the system something new or adjust its behavior in some way. And so a lot of these core capabilities are still missing, and maybe scaling will get us there, but if I had to bet, I think there are probably one or two missing breakthroughs that are still required and will come over the next five or so years,” Hassabis added.
While modern AI systems like GPT-5, Grok 4 and Gemini 2.5 Pro are extremely powerful, winning gold medals at Math Olympiads and coding better than most humans, but they still slip up on some relatively simple tasks. AI models still can’t read analog clocks, draw people with watches on their left hand, or solve many simple riddles. This jagged intelligence, which is extremely capable at some tasks while being extremely weak at others, means that these models can’t yet be considered PhD-level — there are no human PhDs that can’t read clocks. And as Google DeepMind’s Demis Hassabis has pointed out, calling current AI systems “PhD” level might be premature at best, and marketing-driven at worst.