DeepMind CEO Demis Hassabis Warns AI Could Be Misused by States and Bad Actors

Google DeepMind CEO Demis Hassabis has issued a stark AI warning, emphasizing that increasingly powerful AI systems could be misused by bad actors and nation states as the world nears Artificial General Intelligence (AGI).

Speaking at the AI Impact Summit in New Delhi, Hassabis emphasized that AI, while transformative, also carries serious risks if misapplied by malicious actors, including rogue individuals and nation states.

AI Misuse and Emerging Threats

Hassabis flagged the growing threat of AI being exploited in both cybersecurity and biosecurity contexts.

He cautioned that current AI systems, although impressive, are not immune to misuse. Offensive applications could include cyberattacks, large-scale disinformation campaigns, or even facilitating biological risks, areas where robust safeguards are not yet fully in place.

We must ensure that our cyber defences outpace potential attack vectors, especially as AI systems become more capable.

-Hassabis said during his address

AGI Could Arrive Within Five to Eight Years

The DeepMind CEO reiterated previous projections about AGI timelines, suggesting that Artificial General Intelligence could emerge within five to eight years if current trends continue.

Hassabis stressed that AGI represents not only a technological milestone but also a turning point where both the opportunities and risks of AI will dramatically increase. He urged global stakeholders to prepare frameworks that anticipate misuse and ensure safe deployment.

Balancing Innovation With Safety

While acknowledging AI’s immense potential, Hassabis highlighted the need for international cooperation to mitigate risk.

He suggested that companies and governments should:

• Invest in advanced cybersecurity infrastructure
• Develop regulatory frameworks for AI safety
• Ensure ethical standards in AGI research
• Create mechanisms to monitor AI misuse

Without such measures, Hassabis warned, rapidly improving AI systems could be leveraged in harmful ways faster than defensive systems can adapt.

Why This Matters

Hassabis’ warning comes at a time when AI capabilities are accelerating, with tools increasingly capable of generating text, images, and even biological data.

Experts have long cautioned that advanced AI could be weaponized or misused if left unchecked. By highlighting nation-state risks, cyber threats, and biosecurity, Hassabis is underscoring the urgency of global AI governance.

For the tech community, his comments reinforce the need to balance innovation with security and to ensure that society is prepared for AGI-era challenges.