Even Tom Cruise might not have accepted it. Gerard de Graaf was sent a little over two years ago by the European Union (EU) to Silicon Valley with a thorny mission. Explain to American technology companies how to comply with the digital laws promulgated by the Old Continent, from the Personal Data Regulation (GDPR) to the Digital Services Act (DSA), including the recent IA Act. A unique function, both technical – perfect knowledge of these regulations is required – and highly diplomatic. Gerard de Graaf most often interacts with firms like Meta, Apple or Amazon, who spend millions of dollars on lobbying and lawyers in order to water down these texts and escape the sanctions provided for in the event of breaches. Fair play, he recognizes their right to defend their interests: “Speed is essential in digital technology. With network effects, there is always only one big winner. Obviously, all this is done to the detriment of control and security.”
But Gerard de Graaf had until now a powerful ace up his sleeve: a Biden administration with which he had “a good relationship”, recognizes the diplomat of Dutch origin. This notably promulgated an ambitious decree aimed at regulating the artificial intelligence (AI) sector. While, at the same time, pursuing an offensive antitrust policy against “Big Tech”, hand in hand with the EU. The Valley coped with it as best it could. “The tide is turning,” assured de Graaf, a year ago, to Politico. A “Brussels effect” was beginning, according to him, to permeate the American tech oasis. “I was no longer asked why we regulated in Europe, but why the process always took so long,” he remembers, amused.
Patatras. November marks the return of Donald Trump and a whole new team, with a knife between their teeth. Vice President-elect JD Vance is challenging anyone to regulate Elon Musk’s platforms, like his social network X, or risk withdrawing his country’s support for NATO. The newly appointed head of the Federal Communications Commission (FCC), Brendan Carr, is more in favor than ever of high-tech giants aligning themselves with the lowest possible level of moderation. Finally, the shadow of massive deregulation looms larger over AI. Seeming to make Silicon Valley today a land as hostile to Gerard de Graaf as the desert to a penguin.
“Considerable support for regulation in the United States”
Are we wrong? Blue and white polo shirt, slightly disheveled blond hair, the European ambassador living near San Francisco responds to our morning video call with a smile. This is where he often starts his days, in order to debrief as quickly as possible to his Brussels colleagues for whom the sun is already setting. He then goes to his offices, nestled on the 23rd floor of a downtown tower housing the Irish consulate. The Wikimedia Foundation is around the corner, the X headquarters a stone’s throw away. A good hour’s drive south are finally the prestigious headquarters of the tech giants, in Palo Alto, Mountain View and Cupertino. Whatever happens, Silicon Valley remains, according to him, a prime location for anyone interested in cutting-edge technologies. “My role is also to observe the progress made here in AI, quantum, electronic chips…” To separate the wheat from the chaff. “When I took office, the big trend was the metaverse.”
The ambassador, however, kept faith in his first objectives. “We must continue to work with the United States, where there is still considerable support for regulation,” he says, citing the example of the subject of the protection of minors, present in the Digital Services Act ( DSA). Or the fight against anti-competitive practices, at the heart of the Digital Markets Act (DMA). This last text allows everyone to pull each other in the way, to shake up hierarchies, which is in fact very appreciated. “A lot of information that we recover comes from competitors of companies targeted by investigations,” confides Gerard de Graaf, whose small team – which he compares to a start-up – has doubled in volume in two years, passing from 4 to 8 employees. Another good sign.
Concerning AI, in reality, the regulatory issues do not leave political leaders, legislators, or even business leaders indifferent. Elon Musk publicly declared himself in favor of it in the summer. Donald Trump published a strategic decree on this technology in 2020, which included the desire to protect “civil liberties, privacy, American values and the economic and national security of the United States”. AI policy is one of the areas “where there is bipartisan support and shared goals,” notes on the social network LinkedIn, Marc Rotenberg, president of the Center for AI and Digital Policy, an independent non-profit organization. non-profit. De Graaf finally sees growing interest from the general public. “Why do people have confidence when they go to a supermarket or when they take a plane? Because there are food and aviation safety rules. It’s the same in energy, banking , insurance… Everything is regulated While AI is becoming more and more important in our lives, on subjects as essential as health or employment, why would we leave these similar systems unregulated. to black boxes? normal.”
The California effect
The central question remains: how? In the United States, the solution surely lies at the state and not federal level, the ambassador urges. A conviction dating well before the comeback of Trump at the White House. On AI, this year alone, “more than 700 legislative proposals were discussed in 43 separate chambers of the States”, figures Gerard de Graaf. California, a Democratic stronghold, is particularly active. “If you add up all the bills pushed here, you more or less get the scope of the IA Act,” the diplomat continues. For the moment, all these unequal processes in both substance and form are not always successful. If an effort has been made to combat the production of child pornography images created by AI or deepfakes, the regulation of Golden State still has many blind spots. A text aimed at making the use of AI in a discriminatory manner illegal in housing, finance, insurance and health care was notably shelved. Governor Gavin Newsom, in recent months, has also vetoed three other AI attempts including the widely publicized SB1047, which focused on the security of large language models (LLM) with regular audits.
However, a good version could, if it succeeds, have every chance of spreading. “We often talk about a Brussels effect, but there is also a California effect, notes Gerard de Graaf. The European Personal Data Regulation inspired the California Consumer Privacy Act – the CCPA in 2018. Subsequently, the latter was copied by more than fifteen American states.” Which reassures him in the fact of being in the right place. Perhaps more, on the other hand, at the right time.
Trump’s total unpredictability on digital in general, as well as Musk’s growing influence on issues of content moderation – a theme so dear to Europe – remain big question marks. As a good diplomat, De Graaf, 62, prefers to cut it short: “we’ll see.” He temporizes in passing on the strong activity of the EU on the regulatory level in the short term: “We had an intense period of drafting laws. But we are no longer going to chain the “Acts” in this way, we will from now on be more selective.” Enough, perhaps, to calm some ardor. In any case, no one seems more qualified to face the possible storms to come. The discreet special envoy, whose mission runs until 2026, knows the backstage of American political life very well thanks to his role as trade advisor for the EU in Washington, from 1997 to 2001. And even better technology: in 30 years of career at the Commission, de Graaf has worked on the tactical regulation of technologies, Net neutrality, plethora of issues linked to cybersecurity, copyright… “Fortunately that it’s there,” praises Florence G’Sell, visiting professor of private law at the Cyber Policy Center at Stanford University, in Silicon Valley. “It is perfect for explaining what Europe is doing. It is extremely useful in a world where euro bashing is easy.”
.