In the background of trade and tariff tensions, the EU and the US have in recent weeks also escalated their duel in the arena of high-tech, especially AI regulation.
Just days before an important part of the EU’s flagship AI act came into force, Donald Trump issued three AI executive orders as part of his action plan to “cement US dominance” in the field. One of the orders was targeted at “woke AI”.
While the woke AI directive addresses procurement and use by the US federal government, it can also be seen as pushback against the EU’s advancing AI regulatory framework.
From the beginning of August, the bloc’s AI Act applies to “general-purpose AI models with systemic risk.” This includes all the popular well-known AI models, such as ChatGPT, Claude or Gemini, often from US companies, which have another two years to fully comply.
In its guidelines for these platforms, the EU says that it aims to promote innovation and uptake of AI. But it emphasises that this must go hand in hand with “a high level of protection of health, safety, and fundamental rights, including democracy and the rule of law.”
EU-US regulatory tensions
At first glance, the competing EU and US regulatory approaches appear to be part of a more comprehensive commercial and trade policy confrontation.
The EU and US have already clashed on the EU imposition of standards and taxes that extend to US tech companies, as in the Union’s Digital Services Act (DSA) and Digital Services Taxes (DSTs).
But the Trump administration has given these at-heart commercial conflicts a heavily political and ideological load by framing them in terms of censorship, freedom of expression and imposition of woke.
In the field of AI, it is also equating dominance with continued US technological, economic and cultural hegemony and sees the EU’s attempt to regulate as a direct threat to these goals.
Canada already has had to pause a DST on US tech companies or face higher tariffs
Ideology, trade policy and, not seldomly, specific commercial interests are completely enmeshed in the Trump administration, with one being used to enforce or benefit the other.
Canada already has had to pause a DST on US tech companies or face higher tariffs, India now faces tariffs over its dealings with Russia, and countries have been threatened with tariffs over diverse issues such as migration and fentanyl.
The Brussels Effect
The strong anti-EU sentiment in the Trump/Vance presidency underpins much of the dynamic vis-à-vis the bloc.
The 15 per cent tariff deal that Trump and Ursula von der Leyen agreed to in Scotland last month is probably being seen as modest in a US administration that has a visceral dislike for the EU and the multipolar geopolitics and rules-based approach that it stands for.
The EU is a thorn in their side for obvious reasons
Trump, Vance and advisor Stephen Miller share a common trait with many US business magnates, such as Elon Musk and Exxon’s CEO Darren Woods, who has demanded the administration use the trade talks to exempt US companies from the EU’s Corporate Sustainability Due Diligence Directive, which targets human rights and the environment.
They all want as much freedom of action for themselves, whether to perpetuate political dominance, maximise profits, or preferably both.
The EU is a thorn in their side for obvious reasons. Not only is the bloc’s approach, at its best, diametrically opposed to this ultimately arbitrary, self-dealing style of governance. It also has the regulatory heft to make a difference, the so-called Brussels effect that has been amply demonstrated in the past.
Still on the table
Trump’s Woke AI executive order seems at least partly to be a pre-emptive shot to try to make sure that American companies will not be as affected by EU regulation as they have been by the bloc’s General Data Protection Regulation legislation (GDPR).
This was introduced in 2016 and, although not emulated by the US federal government, has had a decisive impact on how American companies deal with online privacy and data protection issues.
While the woke AI executive order is limited to the federal government, it sends a signal to American tech companies that they will be held to the Trump administration’s standards and have to tread carefully when following the EU regulations. Ultimately it could lead to more divergent models for the different markets.
However, this would still not guarantee the US emerging as a regulatory standard setter. On the contrary, the administration is presenting its efforts as anti-regulatory.
This still leaves the door open for anyone interested in implementing responsible AI solutions to depend on EU regulation, from risk-averse large corporations to governments that do not have the wherewithal to devise their own frameworks.
"That is going to be on the table. It wasn't on the table today" - US Commerce Secretary Howard Lutnick
Placing itself at the centre of high-tech regulation has so far not benefited the bloc’s own tech-sector, nor has it created other significant economic gains.
In fact, the European GDPR is often cited as a reason for the lesser availability of data to train AI models on in the EU, although that also has to do with language and market fragmentation and other data constraints. What it has given the EU is something that the Trump administration appears to both abhor and disdain: soft power.
The Trump administration is already girding up for a fight with the EU over the entirety of the tech industry. When Trump and von der Leyen agreed to the trade deal, US Commerce Secretary Howard Lutnick talked about "the attack on our tech companies" as unfinished business. "That is going to be on the table. It wasn't on the table today," he said.
What the EU needs
Clearly, AI regulation is going to be part of that discussion. This is bad news for the EU, just as the linkage between other issues and trade has been problematic for other countries.
EU regulation appears to be more about protecting its principles and standards rather than industry. But for this US administration, that is as much a target as anything else
Purely on its merits, the EU would stand by much of its tech-regulation. GDPR has, according to several studies, significantly increased consumer trust, and the DSA is expected to do the same. Ironically, it’s also US companies that benefit.
With its relatively small tech and AI sector, the EU is largely a consumer rather than a player, much as is the case, for example, in fossil fuel. The US, however, is a supplier and is looking to enable as much access for its products as possible. It has shown that it is not averse to using bullying tactics and tariff threats.
In tech, other than with fossil fuel, there can hardly be any real protectionism. It crosses borders regardless of barriers and regulation, except with draconian regimens such as China’s.
Thus, the entirety of EU regulation will do very little to boost the bloc’s own tech sector, unlike, for example, China’s data localisation laws. The EU would benefit much more from initiatives that increase venture capital mobilisation and reduce market fragmentation.
Thus, EU regulation appears to be more about protecting its principles and standards rather than industry. But for this US administration, that is as much a target as anything else.
