Imagine you are facing a hard problem and, instead of thinking alone, you open an AI tool that helps you explain, compare, and refine ideas in seconds. That is the core of Mo Gawdat’s message when he says people can use AI to increase their effective intelligence. Public sources show that Gawdat, the former Chief Business Officer of Google X, has described AI as something that lets users “borrow” IQ points. The idea matters because it turns AI from a novelty into a practical aid for reasoning, writing, and studying. At the same time, public sources do not clearly confirm that using AI raises a person’s measured IQ score in the formal psychological sense.
Gawdat is speaking to a broad audience, not only engineers. His public talks and official biography present him as a technology executive, author, speaker, and commentator on the social consequences of artificial intelligence. In one published transcript of an April 2025 interview, he said that he goes to his AI and “borrow[s] 40 to 50 IQ points,” framing AI as an amplifier of individual thinking rather than a replacement for the mind itself. In later public remarks around SXSW Sydney 2025, that same theme appeared again, with summaries of his talk describing AI as a way to “borrow” intelligence. For workers, founders, students, and knowledge-heavy professionals, his message is straightforward: the people who learn to use AI well may gain a real advantage in solving cognitive tasks.
In real life, this framing fits best in settings where people already rely on analysis, drafting, and rapid iteration. That includes studying difficult subjects, preparing reports, comparing options, and breaking large questions into smaller ones. Gawdat’s formulation appears most useful when the work requires handling information rather than measuring innate ability. That timing matters because intelligence, as the American Psychological Association explains it, is tied to intellectual functioning and IQ tests compare performance against peers. In other words, there is a difference between performing better with a tool and permanently changing the underlying score that a formal IQ test is designed to estimate.
Public research gives a more grounded view of how AI can help. A randomized controlled trial at Harvard found that an AI tutor in an undergraduate physics course produced higher learning gains, engagement, and motivation than the comparison condition for the material studied. Another randomized experiment with university students found that AI tutor access improved test performance, especially for students who began with less prior knowledge. That suggests a practical mechanism: AI can support learning by giving explanations, feedback, and self-paced practice. Like using a calculator for arithmetic, the tool may improve performance on a task without proving that the user’s underlying capacity has been permanently transformed.
The clearest implication is therefore disciplined rather than dramatic. Gawdat’s public claim is best understood as a metaphor for augmented thinking, not as verified evidence that AI use raises formal IQ. The public record supports his argument that AI can expand what a person can do in the moment. It does not clearly confirm lasting changes in measured intelligence. A sensible next step today is to use AI on one demanding task that benefits from explanation and revision, then verify the result against a reliable source or established course material. That approach follows the evidence: use AI to learn better, while resisting claims that go further than the public facts support.