OpenAI Global Affairs Director Chris Lehane believes that the 2028 presidential election could become a decisive moment for AI. He hopes that politicians will begin to formulate policies to address the concerns of the American public regarding this technology, rather than merely verbally opposing the AI sentiment and the backlash surrounding data centers.
In a new article, Lehane compares the potential of AI to reshape the economy to Franklin D. Roosevelt’s New Deal. The New Deal was a series of federal programs and reforms introduced in the 1930s after industrialization and the Great Depression impacted the old economic order. He points out that Americans today are similarly concerned about losing control over the future of the economy and urges policymakers to find a “convincing third way” that allows people to share in the benefits brought by AI. He lists five key policy areas, covering worker protection, education, AI safety, and data center construction.
At the time of this article's publication, the Democratic Party achieved a stunning victory in the midterms, flipping control of Congress amidst widespread dissatisfaction with living costs. Kate Brennan, who is responsible for AI and corporate power policies, told TechCrunch that the wording of Lehane marks a clear shift in the OpenAI policy agenda; last year, the company's focus was still on accelerating AI development, reducing regulatory barriers, and expanding infrastructure.
As expected, many of the proposals put forward by Lehane encourage the wider adoption of AI and support the infrastructure needed for companies like OpenAI to expand. The initial new policies provided protection for workers and imposed constraints on corporate power. It is not yet clear whether the vision of Lehane will hold AI companies accountable for the economic impacts their technology may have.
"Using the metaphor of new policies, it's about the state's ability to truly benefit the public," said Brennan. "This article comes from a private company that is trying to secure policies for its own interests. At best, this is a misplacement of history; at worst, it is harmful."
Although AI is not the main topic in the mid-term elections, opposition to data centers has become a major issue of concern for both parties in several campaigns. Lehane might be right; in the next two years, AI will likely become an even more important part of the discussions.
Affordability has begun to become one of the key issues in the 2028 elections, as Americans are facing inflation, tariffs, and rising living costs. As voters worry that data centers are driving up electricity costs, that AI is replacing their jobs, and that the economic benefits brought by technology are only flowing to a few wealthy companies, AI is likely to become an increasingly prominent part of this discussion.
Lehane acknowledged many of these concerns, but he hardly proposed any concrete solutions to prevent AI profits from being overly concentrated in the hands of a few individuals. Instead, he mainly portrayed AI as a path to greater prosperity, on the premise that Americans are willing to embrace it.
To take his tax proposals as an example, he suggests providing incentives for workers to learn AI skills, offering entrepreneurs easier access to capital, and exploring new ways of employee stock ownership and profit sharing. However, these are not binding requirements; the latter would ensure that workers share in the wealth created by AI, or require companies to bear the costs of layoffs and job replacements.
"What impressed me is how much of this policy is designed for a future... a future that already presupposes the widespread adoption of AI, and AI will benefit society, yet we haven't seen that yet," said Brennan. "All of this is based on the premise of profiting from the upward gains of AI."
Brennan indicates that a more reasonable approach would be to consider adjusting corporate tax rates, or to provide tax incentives for hiring real people rather than automating jobs, while also trying to prevent companies from accumulating excessive market power.
Lehane focuses on expanding the economic opportunities of AI rather than limiting potential harms, and this extends to data centers as well. He mentioned people's concerns about data centers, while emphasizing the potential for job creation in construction and local tax revenue they can generate. Subsequently, he warned that politicians who support halting the construction of data centers could jeopardize America's leading position relative to China, and even compared this stance to the vote that authorized the Iraq War in 2002.
Lehane also listed child safety and education as a core part of its proposed agenda, emphasizing that AI partners for minors need to meet age-appropriate protective and safety measures. He also warned against excluding AI from the classroom, arguing that this would lead to inequality between wealthy students and "children from rural communities." However, this argument assumes that AI will improve learning, a claim that has not yet been proven. Critics believe it deserves further scrutiny than simply increasing funding or having human teachers provide personalized attention.
Lehane also calls for the establishment of enforceable AI safety standards, independent assessments, public oversight, as well as the ability to restrict AI systems or to shut them down using a "kill switch" – despite recent research indicating that OpenAI and other AI laboratories have not published comprehensive plans for how they would respond if the models lose human control.
Brennan calls these positions "hypocritical" and points out that OpenAI previously opposed legislation aimed at implementing such measures. For example, in 2024, OpenAI opposed California's AI Safety Act SB 1047, which originally required disaster risk assessments and the capability to shut down facilities. Last year, OpenAI also attempted to limit state-level regulation, claiming that it could undermine American innovation. The company has recently made adjustments to some of its positions, supporting several state-level safety laws and expressing its willingness to accept independent evaluators embedded within the company.
Lehane believes that people should retain control over the major decisions made regarding the use of AI, rather than leaving all judgment to machines. However, this also raises a question: in a society where humans are increasingly expected to use AI in education, work, healthcare, and government, what exactly does so-called human control mean?
Throughout the entire article, Lehane portrays AI as a technology that can allow Americans to regain their agency and economic opportunities. However, he also links the "middle class" of ordinary Americans to an economic system that is increasingly shaped by companies such as OpenAI. It is currently unclear whether his policies will truly redistribute power or will instead consolidate power in a more concealed manner.
TechCrunch asked OpenAI for more details about Lehane's proposal and the company's stance, but as of the time of publication, there has been no response from the other party.
