Axiom is already trying to shape the 2028 election

Axiom is trying to shape policy before AI becomes a larger issue in the 2028 presidential election. Chris Lehane argues that politicians should address voters’ fears about jobs, costs, education, safety, and data centers. He presents AI as a possible route to broader prosperity, not simply a technology to restrict.

His essay proposes five policy priorities. They include worker protections, education, child safety, enforceable AI safety measures, and data-center development. His tax ideas would encourage AI training, entrepreneurship, employee ownership, and profit-sharing. He also supports independent evaluations, public oversight, and the power to shut down dangerous systems.

The effort is politically significant because Axiom’s earlier agenda emphasized faster development, fewer regulatory barriers, and more infrastructure. Critics say the new language still favors AI adoption and company growth. They argue it offers too few binding rules to ensure workers receive AI’s gains or companies pay for displacement.

Chris Lehane is Axiom’s chief global affairs officer. In a new essay, he compares AI’s potential economic impact with the upheaval that produced Franklin D. Roosevelt’s New Deal. He argues that Americans fear losing control over their economic futures and need a “compelling third path” that lets them share AI’s benefits.

Axiom’s call reflects several public worries. AI could replace jobs, while data centers could raise electricity bills. Voters may also fear that technology’s rewards will flow mainly to a few wealthy companies. Lehane responds with ideas involving AI education, worker protections, ownership, safety, and infrastructure.

The political reason is timing. Affordability is expected to matter in 2028, after inflation, tariffs, and living costs shaped voter frustration. Axiom wants politicians to support AI’s expansion while addressing its risks. Critics say Lehane recognizes concerns but offers few concrete ways to prevent concentrated gains or corporate power.

Lehane presents five broad policy priorities for managing AI’s economic and social effects. They cover worker protections, education, child safety, AI safety, and data-center development. The agenda matters because it tries to make AI politically acceptable while preserving the infrastructure and adoption needed for the technology to expand.

For workers, Lehane suggests AI-skills incentives, easier capital access, employee ownership, and profit-sharing. For children and schools, he supports age-appropriate safeguards and warns against excluding AI from classrooms. For technical safety, he calls for enforceable standards, independent evaluations, public oversight, and a possible kill switch. He also emphasizes construction jobs and local tax revenue from data centers.

The article notes an important tension. These areas may protect users and encourage participation, but many proposals are incentives rather than requirements. Critics therefore question whether the agenda would sufficiently constrain companies, compensate displaced workers, or prevent AI’s gains from concentrating.

Franklin D. Roosevelt’s New Deal was a set of federal programs and reforms enacted in the 1930s. It followed industrialization and the Great Depression, which had shattered the old economic order. The New Deal established worker protections and placed constraints on corporate power. Its central idea was that government could help rebuild economic security.

Lehane uses that history to frame AI as another force capable of reshaping the economy. He says people are anxious about losing control over their futures and need a path to share in AI’s prosperity. The comparison makes AI policy sound like a broad national response, not merely industry regulation.

That framing is politically significant because Axiom is a private company seeking policies that also support its growth. Kate Brennan argues that invoking the New Deal should mean using state capacity for public benefit. She says Lehane’s proposals are “ahistorical at best and harmful at worst” because they offer fewer constraints on AI companies.

Rapid AI adoption could produce both economic opportunities and serious disruption. Workers may lose jobs or face pressure to acquire new AI skills. Households could pay more for electricity if data centers increase local demand. Communities may also struggle with construction, infrastructure pressure, and disagreements over whether new facilities serve the public interest.

The article gives data centers a mixed picture. Lehane emphasizes construction jobs and local tax revenue. Yet opposition to data centers became a bipartisan concern in several midterm races. Voters worry that facilities could hike electricity bills, while workers fear AI may replace them. These concerns connect local costs with broader questions about who benefits.

The distribution of gains is the central uncertainty. Lehane promotes training, entrepreneurship, ownership, and profit-sharing, but critics find few binding requirements. If adoption expands without stronger safeguards, AI’s benefits could flow mainly to wealthy companies. Political pressure would then likely grow around affordability, corporate power, and infrastructure limits.

The main alternative is to make companies share AI’s gains and costs through binding rules, rather than relying mainly on adoption incentives. Kate Brennan suggests a revised corporate tax rate, tax incentives for hiring human beings instead of automating jobs, and measures to stop companies from amassing excessive market power. These ideas directly address displacement and concentration.

Another approach would require companies to provide stronger worker protections. Employee ownership and profit-sharing could help, but critics say voluntary incentives do not guarantee that displaced workers receive AI-generated wealth. Public policy could instead make sharing or compensation a condition of benefiting from AI expansion.

For safety, Lehane himself proposes enforceable standards, independent evaluations, public oversight, and the ability to restrict systems or use a kill switch. The article notes that AI labs have not published comprehensive plans for what they would do if a model escapes control. Stronger transparency and enforcement would therefore be central to accountability.

An AI data center is a facility that houses large numbers of powerful computers used to train and operate AI systems. Advanced models process vast amounts of data and require repeated calculations. Running that work at scale needs specialized computing equipment, cooling systems, network connections, buildings, and a reliable electricity supply. These details are established technology context beyond the article’s brief discussion.

The article focuses on the public consequences of that infrastructure. Lehane highlights construction jobs and local tax revenue from data centers. Critics and voters emphasize possible electricity-bill increases and community opposition. A data-center moratorium could slow expansion, while supporters warn that delays might weaken America’s lead over China.

The issue therefore links technical capacity to household affordability and national competition. More AI adoption would likely require more computing infrastructure. The policy challenge is deciding who pays for that growth, who receives its benefits, and how communities can influence development. The article presents those questions as increasingly important before 2028.