Eight months before the mid-term elections, artificial intelligence is gradually establishing itself as a political subject in its own right in the United States. Long confined to technological circles and regulatory debates, it is now entering the electoral arena, supported by a coalition of economic and political actors determined to accelerate its deployment, but faced with a public opinion whose mistrust is increasing.
At the heart of this dynamic, David Sacks, White House artificial intelligence advisor, supports the emergence of a political infrastructure dedicated to promoting a pro-innovation agenda. The launch of the Innovation Council, a structure financed by undisclosed funds, is part of this logic. Led by Taylor Budowich, a former close friend of the president, the organization intends to directly influence the 2026 elections by supporting candidates aligned with the administration’s technological strategy.
An assumed political strategy around AI
Since the start of his second term, Donald Trump has stepped up initiatives to accelerate the development of artificial intelligence. A presidential decree aimed in particular to limit the capacity of States to adopt their own regulatory frameworks, in a logic of federal centralization. At the same time, the administration actively supports the construction of data centers and emphasizes the need for the United States to maintain a strategic advantage against China.
This orientation is part of an assumed industrial approach, where AI is seen as a lever of economic and geopolitical power. It finds an echo in business circles, where several figures from tech and venture capital, like Marc Andreessen or Greg Brockman, have committed significant financial resources to support candidates in favor of limited regulation.
In total, several hundred million dollars could be mobilized in the midterms, demonstrating a desire to structure a political environment conducive to the rapid expansion of AI.
Public opinion in sharp swing
This strategy comes against a backdrop of rapid deterioration in Americans’ perception of artificial intelligence. According to a Quinnipiac survey carried out in March, 55% of respondents believe that AI will have more negative than positive effects on their daily lives, an increase of 11 points in one year.
Concerns primarily relate to employment: 70% of those surveyed anticipate a reduction in professional opportunities linked to automation. Conversely, only 7% envisage a positive impact on the labor market. Concerns also extend to education, with nearly two-thirds of respondents believing that AI could deteriorate the school system, as well as to military uses, which the majority view with reservations.
Beyond uses, it is now the very infrastructure of AI that is giving rise to opposition. Nearly 65% of Americans say they are against the construction of data centers in their community, citing energy costs, water consumption and local nuisances in particular. This point of friction, long peripheral, tends to emerge as a local political issue capable of directly influencing the elections.
A dissociation between adoption and trust
The contrast between uses and perceptions accentuates this tension. More than one in two Americans say they use artificial intelligence tools, while a large majority express some form of distrust towards them. This dissociation, already observed in other digital sectors, takes on a particular dimension here due to the economic and social implications of technology.
In this context, the warnings coming from certain industry players are helping to fuel the debate. Dario Amodei recently spoke of a potentially “painful” impact on the labor market, emphasizing the scale of the transformations to come.
AI as a new electoral fault line
As the midterms campaign intensifies, artificial intelligence tends to structure an unprecedented fault line. On the one hand, a coalition of political and economic actors is calling for rapid acceleration, in the name of competitiveness and technological sovereignty. On the other hand, a growing part of public opinion expresses reservations, even opposition, regarding the anticipated consequences of this transformation.
In this context, the 2026 elections will determine to what extent this pro-AI agenda can impose itself politically, or if it must deal with an increased demand for supervision and regulation. Behind the technical debates, it is the capacity of institutions to arbitrate between innovation and social acceptability that is posed.
