Trump entrusts AI to the head of the intelligence services: Clayton is the new ‘czar’ of Super Intelligence

Trump has been referring to it as a ‘race’ since at least July 2025, when the White House published “Winning the AI Race», over ninety federal initiatives to accelerate innovation, build the necessary infrastructure and bring American technology and standards to international markets. ‘Whoever has the largest AI ecosystem will set global AI standards’, the plan states. Whoever controls the largest ecosystem will set global standards and, according to the White House, reap economic and military benefits as a result.

To build that ecosystem, the models developed by OpenAI, Google, Anthropic or xAI are not enough. We need semiconductors, data centres, electricity, networks, capital and access to markets. The plan tasks the Departments of Commerce and State with working alongside industry to offer US ‘full-stack’ packages to allies. That is to say, hardware, models, software, applications and standards.

‘If we don’t, then China will easily catch up with us in the AI race’, ‘if we don’t do this, China will easily catch up with us in the race for artificial intelligence’, Trump warned two years ago whilst preparing his offensive against state laws deemed too restrictive. The order signed a few days later instructed the Department of Justice to set up a task force to challenge regulations incompatible with federal policy and called for a national framework that was ‘minimally burdensome’. Washington aimed to prevent rules that varied from state to state from imposing a growing web of obligations on companies, just as the White House was asking them to step up their efforts.

The President recently claimed that the United States is about eighteen months ahead of China in the field of artificial intelligence and rejected the idea of slowing down its development for security reasons. To do so, he said, would mean risking the loss of that lead over Beijing.

On 5 June, Trump ordered the accelerated deployment of artificial intelligence ‘in intelligence activities and military operations’. The presidential memorandum on national security calls for facilitating access to the most advanced models and for building highly secure computing infrastructure capable of running them on a large scale. The recipients include the Director of National Intelligence, the Director of the CIA, the FBI and senior Defence officials. The stated aim is to maintain US ‘technological superiority’ over adversaries and strategic competitors.

When it comes to risks arising within companies, Trump has so far taken a different approach. Nvidia, SpaceX, OpenAI, Anthropic, Meta and Google have signed an agreement at the White House committing them to strengthening internal controls, subjecting their systems to independent audits and protecting them from unauthorised access and attacks. Participation is voluntary and there are no specific penalties for those who fail to meet the commitments. The President described the agreement as ‘morally binding’.

73 per cent of Americans surveyed in a Reuters/Ipsos poll in September believed that technology companies were not doing enough to prevent artificial intelligence from causing serious harm to society. Fifty-five per cent thought it would be a good idea to slow down its development in order to work on safety. The Oval Office, however, continues to press ahead, leaving a substantial part of the oversight to the companies and now entrusting the Super Intelligence Force with the task of assessing whether the tools available to the government are sufficient.

Meanwhile, the race that Trump wants to win is no longer called artificial intelligence. Since 29 September, in US government documents, ‘Artificial Intelligence’ has been replaced by ‘Super Intelligence’. Executive Order 14434 requires federal departments and agencies, to the extent permitted by law, to use ‘SI’ instead of ‘AI’ in official correspondence, on websites and in reports. Seventy years after the coining of the term ‘artificial intelligence’, the White House argues, the capabilities achieved by the technology have now rendered it inadequate.

However, the technology that US law associates with those terms has not changed. The same order stipulates that ‘Super Intelligence’, at least for the time being, continues to refer to systems already covered by the federal definition of ‘artificial intelligence’. Nor does the order claim that the United States has achieved that ‘superintelligence’ which, in scientific discourse, generally refers to systems capable of surpassing human cognitive abilities. The definition must catch up with the name. Within sixty days, the Presidential Adviser on Science and Technology must propose a new definition and indicate what legislative changes are necessary to incorporate it into federal law.

For the time being, Washington has a Super Intelligence Force, a Director of National Intelligence tasked with leading it, and a government that is required to refer to what the law continues to define as ‘artificial intelligence’ as ‘Super Intelligence’.