In Agenda Digitale, Alessandro Longo and Nicoletta Pisanu examine what the US midterm elections of 3 November 2026, which gave the Democrats control of both the House and the Senate, could mean for digital policy and artificial intelligence. The piece gathers the views of several experts, including Rocco Panetta.

The picture that emerges is one of possibly closer political scrutiny of AI: from January 2027 the new Congress will be able to summon the heads of the large technology companies and request documents on safety, competition and their dealings with the administration. President Trump’s veto power, however, makes a federal regulatory shift far from certain. A second front concerns data centers, the subject of dozens of local ballot measures and promised moratoriums, with possible effects on the cost and timing of AI services that are also used in Europe.

In his contribution, Rocco Panetta acknowledges that the vote opens up new scenarios in the international debate on digital policy, the data economy and the development of AI, but warns against easy expectations: “anyone who expected a clear and swift reversal of course is mistaken”. In his view, the current administration has departed from its predecessors more in form than in substance.

Panetta traces the transatlantic relationship on data back to Directive 95/46, which shaped the European data market some thirty years ago: since then, regardless of who was in power, the United States has kept an approach opposed to the European one. That divide gave rise to the Safe Harbor agreement, negotiated when Stefano Rodotà chaired the group of European data protection authorities, and later to the Privacy Shield, both designed to harmonise EU–US data flows and both eventually struck down.

The gaps, Panetta concludes, are clear and are set to remain, especially on AI regulation: without a strong commitment from the US Presidency, a treaty on the responsible use of AI and the non-proliferation of robotic weapons looks hard to achieve in the medium term.

The article rounds off its analysis with a comparison between the EU AI Act and US initiatives, from California’s SB 53 to the FRONTIER Act, and with a reflection on Europe’s technological sovereignty.