The Future of AI: With Sir Demis Hassabis & Dame Wendy Hall

If I told you that the future of artificial intelligence could be decided in the next four years, would that sound like provocation to you? Yet this is the estimate of Demis Hassabis, founder of DeepMind: according to him, the probability that a form of general artificial intelligence, capable of the same cognitive abilities as a human being, will emerge by 2030 is 50%. But the real question is not “when,” it is “who will be prepared for it – and who will be left watching from the sidelines.” The common view, especially in Europe, is that we can simply regulate AI, perhaps leaving the hard part – inventing, building, leading – to the United States and China. But this is a dangerous illusion: as Hassabis says, “you can't be world champions just in regulation.” Either you have the technology and a say in the matter, or others will write the rules. Take Dame Wendy Hall, a pioneer of web science: as early as 2017, she helped the British government draft one of the first national strategies on AI. At the time, the United Kingdom ranked third in the world, behind only the United States and China. It seemed like a race that could be won. But today, Hall says, the strategy has fallen apart: there is no longer a go-to person for AI expertise, the independent oversight body has disappeared, while countries like Singapore have taken the British “playbook” and surpassed it. And the result is clear: the United Kingdom is slipping back and risks losing the race precisely in terms of skills, training, and the ability to generate genuine innovation. Hassabis is even more forthright: a revolution in schools is needed. We cannot continue to teach as if AI did not exist, or worse, hoping that it will “go out of fashion.” Today, students use chatbots and LLMs to do their homework, often to avoid the effort, but the education system remains stagnant. According to him, the model needs to be flipped: basic concepts and theory are studied at home, perhaps with personalized AI that adapts the level of difficulty, while school becomes the place for relationships, projects, and human skills – creativity, judgment, and teamwork. It’s a bit like a turbocharged Montessori high school, where the teacher is no longer just a presenter of slides but guides students through hands-on experiences. And what about personalization? It’s not science fiction: back in the 1980s, Wendy Hall was already writing code for intelligent tutoring systems, and today she is working to provide an “AI license” to all students at the University of Southampton, even those who don’t study computer science. On the other hand, the game is also about who gets their hands on the key technologies. DeepMind was founded in London, in part to show that cutting-edge science can be done outside Silicon Valley. Hassabis says that in 2010, he could hire top talent “for £50,000 a year,” figures that are unthinkable today. Now, major tech companies have opened offices in London, bringing investment and creating an ecosystem where talented people stay and spawn other startups. And this is where the twist comes: if you thought that all this was just a matter of algorithms and patents, you're missing a piece of the puzzle. Behind every AI breakthrough lies a political, social, and human choice. Hall points out that diversity – in terms of gender, age, and background – is the real key to preventing AI from being built by a male, Western monoculture. And she warns: “AI is too important to leave it to men alone.” Genuine involvement of all voices is needed, not least because the risks are enormous: on the one hand, AI can accelerate scientific discovery, help medicine, and revolutionize energy; on the other, if left unregulated or overseen only by those with economic interests, it can be used for destructive purposes. Hassabis speaks openly about “biorisks,” the danger that generative models could be manipulated by hostile actors. That is why, according to both of them, an independent international agency is needed, a new type of governance in which governments, companies and civil society collaborate – not only to set standards, but to ensure that technology serves everyone, not just a few. And this is a race against time: if AGI really could arrive within a few years, institutions are dramatically behind. The paradox? All stakeholders, from companies to states, have an interest in avoiding disasters – “no one wants a catastrophe to happen, not even the most capitalist,” says Hassabis – but political fragmentation and the logic of the “prisoner's dilemma” threaten to bring everything to a standstill. And if you think the problems are merely technical, you’re mistaken: Hall insists that the real question is what kind of society we want to build and what meaning we ascribe to work, well-being, and equality in the age of AI. If we leave the conversation only to experts, we miss the opportunity to collectively shape the future. One thing is certain: the future of AI is not written, but the window to influence it is closing fast. In short, you can no longer afford to be just a spectator. If you think the game is for a few insiders, you are excluding yourself from the table where the next decade – and perhaps more – will be decided. No one can be just a regulator: either you build, or you suffer. If this discussion has given you a new idea, you can note on Lara Notes that this perspective is now part of how you think with I'm In. And if tomorrow you happen to talk about it with someone – whether it’s about education, governance, or the future of AI – on Lara Notes, you can use Shared Offline to remind yourself who was part of that conversation that mattered. This Note comes from an evening at the Worshipful Company of Information Technologists and saves you 78 minutes.