OpenAI CEO Sam Altman may be softening his stance on the prospect of AI-driven job losses, but Anthropic CEO Dario Amodei is moving in the opposite direction. Amodei is doubling down on his warning that artificial intelligence could profoundly disrupt labor markets and is calling for governments to take a far more active role in cushioning the blow and sharing the gains more fairly.
In a recent essay titled “Policy on the AI Exponential,” Amodei contends there is a serious chance that artificial intelligence will trigger large-scale and enduring job losses, even under relatively careful deployment. He argues that this risk is not merely a byproduct of adoption but is closely tied to AI’s core design goal: replicating human cognitive capabilities at scale.
The concern, he writes, comes from AI systems being engineered to execute tasks that previously depended on human reasoning, judgment, and expertise. As these systems advance, they could increasingly substitute for workers in diverse white-collar and professional roles, reshaping employment patterns far beyond manufacturing and routine services.
Amodei’s warning is not new, but his language is growing more urgent. Earlier this year, he cautioned that AI might wipe out as much as half of all entry-level white-collar positions within the next five years. In that scenario, unemployment could rise to between 10% and 20%. He also pressed corporate executives and policymakers to stop minimizing the dangers tied to rapid AI deployment and to confront the implications head-on.
The new essay builds on that message, arguing that AI development is moving considerably faster than regulatory systems, labor-market policies, and social safety nets can keep up. Amodei says society is approaching what he calls “Powerful AI”—systems able to perform many cognitive tasks at or above the level of human experts across multiple domains.
Once those systems arrive, he argues, the economic and social effects could be both sweeping and persistent. For that reason, Amodei insists that policymakers must act before those systems are fully entrenched, rather than waiting until the consequences are plainly visible in unemployment data or wage trends.
Amodei also draws a sharp contrast between current AI advances and earlier technological shifts. Past industrial revolutions, he notes, mostly automated physical tasks or eventually generated substantial new categories of work that absorbed displaced workers. By contrast, advanced AI threatens to automate the very cognitive functions that underpin many professional and knowledge-based careers.
That difference, he suggests, raises the prospect of sustained job displacement rather than the shorter-term dislocations that have historically accompanied innovation. If AI can reliably replicate complex decision-making, analysis, and creative problem-solving, entire segments of the labor market could shrink without obvious replacement roles.
Amodei’s stance sets up a growing divergence within the AI industry’s leadership ranks. While Altman now appears more inclined to argue that earlier fears of mass unemployment may have been overstated, Amodei is explicitly warning that the worst-case scenarios remain plausible. For governments, the message is clear: treat AI’s labor impact as a central policy challenge, not an afterthought.
Read Article: Sridhar Vembu criticizes Salesforce over pricing tactics in new public dispute
