DOES AMERICA NEED A SECRETARY OF ARTIFICIAL INTELLIGENCE?
Lessons from earlier industries and warnings from AI leaders.
E-ISSN: 2354-4481
By King Eze
[email protected]
Artificial intelligence is no longer merely a technology product. It is becoming an economic engine, military capability, medical assistant, educational instrument and potential instrument of cyberwarfare.
The recent promise by Anthropic CEO Dario Amodei to grant independent safety evaluators extensive access to his company’s systems raises a fundamental question: Should public safety depend principally on voluntary commitments made by AI companies?
America needs cabinet-level leadership for artificial intelligence, but the country must learn from the regulation of earlier transformative industries.
Lessons from previous industries
When aviation expanded, America did not leave every airline to determine its own safety standards. Federal institutions established rules for aircraft certification, pilots and air-traffic management.
After the growth of nuclear technology, Congress eventually separated promotion from safety regulation. The independent Nuclear Regulatory Commission became responsible for nuclear safety because the institution encouraging an industry should not also be its only watchdog.
The financial collapse of 1929 similarly demonstrated that markets could not depend entirely on voluntary corporate restraint. Congress created the Securities and Exchange Commission in 1934 to protect investors and restore confidence.
America also established the Department of Energy in 1977 to coordinate responsibilities previously scattered across different agencies. Following the September 11 attacks, the Department of Homeland Security brought numerous security functions under unified leadership.
The lesson is clear: when an industry becomes powerful enough to affect the entire nation, government eventually creates institutions capable of coordinating its opportunities and controlling its dangers.
AI leaders are requesting oversight
Calls for regulation are not coming only from politicians.
During his May 2023 Senate testimony, OpenAI CEO Sam Altman supported the creation of a federal agency capable of licensing and monitoring the most powerful AI models. The hearing record shows that industry leaders and experts were already asking Congress to establish rules for advanced systems. U.S. -Senate hearing record.
Microsoft President Brad Smith compared AI safeguards to the braking mechanisms used in elevators, buses and high-speed trains. His central argument was simple: “Whenever you have something that fundamentally can do good, but could also…do harm, you put a braking mechanism in place.” He also supported a new agency and licensing system for the most powerful models.
-CBS interview with Brad Smith.
Anthropic’s Dario Amodei has gone further. He argues that when an AI model presents extreme risks, “the government should have the legal authority to block or deter its deployment.” Importantly, he also warns that regulation must prevent government overreach and protect innovation.
-Anthropic’s AI policy proposal.
Geoffrey Hinton, one of the pioneers of modern AI and a Nobel laureate, has warned that companies alone cannot carry the burden. In his words, “government’s the only people who are powerful enough to deal with these large companies.”
-Nobel Prize interview with Geoffrey Hinton.
These voices do not agree on every policy. Nevertheless, they share one conclusion: advanced AI cannot be governed through corporate goodwill alone.
What a Secretary of AI should do.
