Google's Sundar Pichai says AI could do his job one day (www.businessinsider.com)

Google CEO Sundar Pichai told the BBC that AI might one day be able to perform even his job, arguing rapid advances—particularly agent-style models that can act autonomously on users’ behalf—will enable systems to carry out increasingly complex tasks within months, not years. He said the next 12 months will show evolution toward agents that can manage multi-step workflows, a technical trajectory that echoes comments from OpenAI’s Sam Altman about AI potentially outperforming human CEOs. Pichai framed this as both an opportunity and a disruption: AI can free people to focus on higher-level work and create new roles, but it will also displace some jobs and require broad adaptation. For the AI/ML community, Pichai’s remarks underscore two practical implications: accelerating research and engineering priorities around robust, safe agent architectures and growing pressure to address deployment risks, governance and workforce transition strategies. Labor-market signals—Indeed’s recent study highlighting software engineers and product managers as especially exposed—reinforce the need for tooling, education and policy frameworks that help people adopt AI capabilities responsibly. Pichai urged the next generation to learn these tools, implying that skill shifts and human–AI collaboration will be central to how benefits and harms unfold.