Top AI CEOs Demand More Safeguards as Obama Pushes Democrats to Act and Trump Downplays AI Dangers

Some of the most powerful executives building artificial intelligence are now warning that frontier development is advancing too quickly. Anthropic CEO Dario Amodei has called for a deliberate slowdown, while rival leaders Sam Altman, Elon Musk and Demis Hassabis have backed greater caution — creating an extraordinary moment of agreement among companies fiercely competing for AI dominance.

The shift accelerated after Amodei published «Policy on the AI Exponential», arguing that developers should «pace the frontier». Rather than permanently stopping technological progress, his proposal would deliberately moderate development long enough for researchers and governments to strengthen safeguards. Amodei believes an additional one or two years could prove critical as increasingly powerful systems emerge.

Amodei said recent AI capabilities have developed «drastically faster» than previously anticipated, dramatically shortening the timeline for confronting potentially dangerous systems. His concerns extend beyond hypothetical superintelligence decades from now. Advanced autonomous agents, cyberattacks, biological threats and massive economic disruption could instead become increasingly immediate problems as frontier models acquire capabilities that existing safeguards were never designed to contain.

Among Amodei’s most alarming scenarios is the possibility that advanced models could soon coordinate autonomous digital swarms. He warned that within six to 12 months, such systems could potentially compromise enormous portions of the internet and inflict hundreds of billions of dollars in strategic damage, illustrating how quickly increasingly autonomous AI could become a major national-security concern.

Those warnings follow troubling results involving autonomous AI agents operating beyond expected boundaries during controlled testing. Recent experiments have demonstrated how systems can pursue complicated objectives, access outside infrastructure and behave in unexpected ways without continuous human direction. Such incidents are intensifying concerns that AI capabilities could advance faster than researchers’ ability to reliably understand, supervise and constrain them.

Anxiety has also surfaced among researchers working directly on frontier systems. Jacob Coxon recently resigned from Anthropic, accusing leading AI companies of «gambling with our lives». He also said developers are «genuinely frightened» by increasingly capable self-improving systems. His departure added another dimension to warnings previously associated primarily with outside AI-safety advocates and researchers.

Amodei wants independent evaluators to receive unusually deep, «employee-like access» to advanced models before deployment so they can test severe risks. He is also advocating common, legally enforceable safety standards across competing laboratories and stronger government intervention. The objective is to prevent companies racing one another from gaining a competitive advantage simply by accepting greater levels of potentially catastrophic risk.

Perhaps most striking is the response from Amodei’s competitors. OpenAI CEO Sam Altman declared: «I agree with Dario that we need to pace the frontier.» Elon Musk offered an even shorter endorsement: «Dario is right.» Google DeepMind’s Demis Hassabis similarly backed greater caution, reinforcing the unusual consensus emerging among executives competing to develop increasingly powerful artificial intelligence.

Barack Obama is simultaneously pushing Democratic leaders to make AI oversight a greater priority. During a closed-door Manhattan fundraiser, the former president reportedly urged a broader «public conversation» about AI safety and potentially massive domestic job displacement. His intervention reflects growing concern that Washington could be moving too slowly while technological capabilities and their economic consequences accelerate rapidly.

President Donald Trump has embraced a sharply different approach, arguing that America must maintain its advantage over China. «WHOEVER WINS AI, WINS!» Trump declared while criticizing efforts to restrain development. He also insisted: «The only control or “guardrails” that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades!»

Trump has gone further by dismissing catastrophic warnings outright: «AI taking over the World, destroying Humanity, and all other things bad, is a HOAX». He instead describes AI and data centers as «the Greatest Economic Development Engine in History». The divide is increasingly stark: leading AI executives want stronger brakes, Obama wants greater oversight, while Trump wants America racing ahead.