World losing pace to keep up with potential AI risks, expert warns

World losing pace to keep up with potential AI risks, expert warns
—————————–
A leading AI safety researcher has warned that the world may not have sufficient time to prepare for the risks posed by rapidly advancing artificial intelligence systems. David Dalrymple, a programme director at the UK government-backed Advanced Research and Invention Agency (Aria), said the pace of AI development could outstrip efforts to ensure its safe use.

Speaking to The Guardian, Dalrymple said concern was warranted as AI systems approach the ability to perform a wide range of human tasks more efficiently than people. He warned that such capabilities could undermine economic stability, security and governance if not properly controlled.

Dalrymple highlighted a growing gap between public-sector understanding and the expectations of AI companies regarding imminent technological breakthroughs. He said governments should not assume advanced AI systems are reliable, noting that economic pressures may limit the time available to develop robust safety science. Instead, he argued, priority should be given to mitigating risks and controlling potential harms.

His comments come as the UK’s AI Security Institute reported that advanced AI capabilities are improving rapidly, with performance in some areas doubling every eight months. The institute said leading models can now complete complex tasks autonomously, though it stressed that worst-case scenarios such as uncontrolled self-replication remain unlikely in real-world conditions.