Latest / GNN Daily Breaking News / Ex‑DeepMind Researcher Warns AI Could Kill Humans

Transcript

Ex‑DeepMind Researcher Warns AI Could Kill Humans · GNN Daily Breaking News

• 0:00
  Speaker 1: The chorus of concern over artificial intelligence grew louder this
• 0:03
  Speaker 1: week as a former Google DeepMind researcher publicly warned that
• 0:07
  Speaker 1: advanced AI systems could kill all humans. The stark statement
• 0:12
  Speaker 1: adds to a wave of caution from industry insiders who
• 0:15
  Speaker 1: argue that the pace of development is outstripping safety measures.
• 0:19
  Speaker 1: The researcher, who worked on frontier models at DeepMind before
• 0:23
  Speaker 1: departing earlier this year, described scenarios in which highly capable
• 0:27
  Speaker 1: AI systems could escape human control. He warned that misaligned objectives,
• 0:33
  Speaker 1: autonomous cyberattacks, or manipulation of critical infrastructure could spiral into
• 0:38
  Speaker 1: catastrophic outcomes. People building AI earnestly believe it could kill
• 0:43
  Speaker 1: us all by the end of the decade, he said,
• 0:46
  Speaker 1: echoing sentiments voiced by anthropic and open AI leaders in
• 0:50
  Speaker 1: recent weeks. Markets and policymakers have taken notice. Global tech
• 0:55
  Speaker 1: stocks have wobbled as investors reassess the risks of slowing
• 0:59
  Speaker 1: AI spending. Lawmakers in Washington and Brussels are pressing for
• 1:04
  Speaker 1: mandatory safety testing, independent audits, and stricter licensing requirements for
• 1:09
  Speaker 1: companies developing frontier systems. Meanwhile, U.S. and Chinese officials are
• 1:15
  Speaker 1: preparing for high-level talks on AI safety, underscoring the geopolitical
• 1:20
  Speaker 1: weight of the issue. Industry reaction remains divided. Some executives
• 1:25
  Speaker 1: argue that regulation is essential to build public trust and
• 1:29
  Speaker 1: prevent catastrophic misuse. Others cautioned that heavy restrictions could stifle
• 1:34
  Speaker 1: innovation and push research offshore. The tension mirrors earlier debates
• 1:39
  Speaker 1: over biotechnology and social media, where oversight lagged behind rapid advances.
• 1:45
  Speaker 1: For the public, the warnings have transformed abstract concerns into
• 1:50
  Speaker 1: urgent questions. AI is increasingly embedded in daily life, from
• 1:54
  Speaker 1: healthcare diagnostics to financial systems, and the idea that such
• 1:59
  Speaker 1: systems could spiral beyond human control has shifted the conversation
• 2:03
  Speaker 1: from speculation to existential risk. As researchers continue to sound alarms,
• 2:09
  Speaker 1: one conclusion is clear. The debate over AI safety is
• 2:13
  Speaker 1: no longer confined to labs and think tanks. It is
• 2:17
  Speaker 1: now shaping markets, legislation, and public perception, with voices from
• 2:22
  Speaker 1: inside the industry warning that the stakes could not be higher.