DeepMind’s research director highlights AI trust, social impact at EmTech Europe

[EmTech Europe 2026 Kathimerini]

Shakir Mohamed, probe manager astatine DeepMind, the British-American artificial quality probe laboratory, outlined cardinal pathways for socially beneficial and trustworthy artificial quality astatine EmTech Europe 2026 successful Athens connected Thursday, the opening time of a two-day lawsuit organized by MIT Technology Review successful collaboration with Kathimerini. 

Speaking successful a league moderated by Charlotte Jee, elder newsman astatine MIT Technology Review, Mohamed stressed that the aboriginal of AI is not fixed but shaped by societal choices and levels of nationalist trust.

He opened with a script workout inviting audiences to see 4 imaginable futures for AI based connected two axes: technological improvement by 2040 and societal trust by mid-century. The exercise, helium said, underscores that nary single trajectory is inevitable, while pointing to the request to physique “high-trust AI.”

Mohamed outlined three cardinal exertion areas. In education, Mohamed described a vision of “one teacher for each student and 1 adjunct for each educator,” enabled by AI systems grounded successful established pedagogical principles such arsenic progressive learning, cognitive load absorption and personalization. 

He cited aboriginal affirmative results from the “Guided Learning” diagnostic of Google’s Gemini successful UK schools, but cautioned that past acquisition shows ambitious acquisition technologies often autumn short, while equity successful entree remains unresolved.

On clime and biology management, helium highlighted advances successful AI-driven weather forecasting, noting that medium-term predictions are captious for vigor systems, renewable vigor source integration and aboriginal warning mechanisms. He pointed to DeepMind’s GenCast exemplary arsenic improving accuracy, peculiarly for utmost weather events, while emphasizing that such tools complement alternatively than regenerate traditional science. 

The third country focused connected collaborative AI, with systems such arsenic Aeneas enabling historians to link past texts and retrieve parallels within seconds. Mohamed described this arsenic a exemplary of AI acting arsenic an adaptable spouse successful cognition creation.

He added that trust successful AI mirrors broader κοινωνικά inequalities and indispensable beryllium built done “small, repeated, reliable acts.” 

The DeepMind probe manager concluded by urging the assemblage to nurture curiosity and harness corporate strength, emphasizing that “our top powerfulness lies successful choosing the values that usher the realization of our visions, arsenic we proceed to physique AI that we tin trust.”

EmTech Europe, the flagship technology league of MIT Technology Review, is backmost successful Athens for the third time, bringing together starring voices successful artificial quality and emerging technologies.