The new AI Action Plan of the US president abandons all caution: more defence, less regulation, full alliance with Big Tech. For Europe, it is time to decide whether to remain bound by its regulations or enter the global industrial game. by Andrea Monti – Initially published on Italian Tech – La Repubblica
The meaning of the document just released by the White House on the American strategy for AI is all in the title: “Winning the Race”.
In no uncertain terms, the Trump administration goes beyond the Biden presidency’s AI executive order issued on 14 January 2025 and then repealed by the current administration, which more mildly spoke of “growing US leadership in AI infrastructure”, and clearly defines objectives and tools to achieve “victory”.
The autonomy of the manufacturing chain
Clearly mindful of the chip shortage, the rare earth crisis and the renewed awareness of the problems of a much too long supply chain, US policymakers have first and foremost focused their attention on creating a “short” manufacturing chain, a local workforce and research that attracts foreign talent. This is no different from what has already been done in other industrial sectors to forcibly import foreign production and skills into the US.
Sovereignty over datasets and the role of science
Similar to what was established a couple of years ago by the Chinese government, the US also attaches particular importance to datasets, especially those intended for scientific research.
In this sense, therefore, the instrumental role reserved for AI is clear, whose development is not an end in itself, but also (and above all) serves to support other sectors critical to the US administration, from scientific research to government administration and defence.
The fusion of AI, genetic manipulation and biosecurity
Almost as an afterthought, the document makes explicit reference to the importance of strengthening research into biosecurity.
This could mean, as the document states, putting AI at the service of genetics to cure diseases, but also, one might infer, using AI to find countermeasures to biological agents, i.e., reopening Pandora’s box of bacteriological weapons. It is impossible not to hear, in this choice, an echo of the issues surrounding the causes of the pandemic.
Market protection and knowledge’s control
Following a fairly well-established technology policy, especially in the arms sector, the document states in no uncertain terms that US technologies cannot be made indiscriminately available (indeed, they must be scrupulously protected, even from internal industrial espionage) while at the same time strengthening the alliance with Big Tech, which is increasing its role as a strategic partner of the government.
Regulatory simplification
The driving force behind the Trump administration’s strategy is a clear choice on regulation: few rules, easily enforceable.
Contrary to the approach taken by the EU with the AI Regulation and related implementing acts, such as the Guidelines on the scope of the obligations for general-purpose AI models, the US is clearly aware of the need to take sides on the role of AI companies: to prioritise the ability to develop increasingly efficient tools, without the burden of unnecessary regulatory red tape.
“I is too important, the document states, to be bogged down in state or federal bureaucracy at this early stage. The federal government should not allow public funds to be directed to states that enact burdensome laws, but without interfering with the right of states to pass prudent laws that do not impose unnecessary restrictions on innovation.
The comparison with the EU
It may be a coincidence that the new US strategy for AI was released at the same time as the tariff negotiations, but this document is certainly destined to play a role in the ongoing diplomatic talks.
After the GDPR, the AI regulation also risks becoming a bone of contention when, from the US perspective, it prevents or severely restricts the spread of American technologies.
These restrictions are not only relevant in terms of trade or maintaining technological control over EI member states, but also affect the overall strategy to counter Chinese technologies.
The EU has not yet taken a clear position towards Beijing on a number of issues, to which AI and the technologies that enable it can now be added. It is therefore not unreasonable to think that the US may also issue an ultimatum to Brussels in this area, whose much-invoked “bazooka” — “attacking” Big Tech — seems increasingly unlikely to be used in the field.
Conclusions
The US administration’s choice is clear: on the one hand, it is pushing for rapid deregulation and militarisation of innovation, while on the other, it is relegating any consideration of rights to the background. In geopolitical terms, therefore, the new American strategy on artificial intelligence is not only a gigantic industrial plan, but above all an assertion of power to put the United States back at the centre of a new geography of knowledge.
Faced with this stance, if the European Union really wants to talk about digital sovereignty, the time has come to move beyond the bureaucratic and dirigiste regulatory phase and equip itself with an industrial strategy on AI that is comparable in ambition to that of the United States. Otherwise, regulations will only serve to slow down the spread of other countries’ technologies, not to generate those of the future in Europe.