第一财经

American AI company invests tens of millions of dollars in the mid-term elections of 2026: What are their intentions?

原文:美国AI公司豪掷数千万美元押注2026中期选举, 有何图谋?

Summary of Key Points

During the 2026 U.S. midterms, AI companies and their associated super political action committees (super PACs) spent over $200 million to support candidates and lobby for policies, with the aim of influencing AI regulation legislation. Two major AI super PACs achieved remarkable results, with most of the candidates they supported winning. However, there is a disagreement on whether AI regulation should be unified at the federal level or left to the states. Additionally, it was rare for members from both parties to agree on this issue, as they are concerned that AI chatbots could provide voters with incorrect information and interfere with the fairness of elections, calling on relevant agencies to address these concerns.

I. AI Companies Spending Money on Elections: It's About "Policy Investment"

Why are AI companies willing to spend so much money on elections? Simply put, it's about using money to influence policy—policy is crucial for the survival and development of AI companies. The severity of regulations, whether data centers can be built, and the openness of the innovation environment directly determine their ability to make more money.

For example, in 2024, the cryptocurrency industry spent $200 million to support candidates, which led to the passage of stablecoin legislation and progress on digital asset regulation, essentially "recouping their investment." Now, AI companies are doing the same, hoping to influence future AI legislation in their favor by supporting candidates they approve of (e.g., advocating for less stringent regulations or granting approval for data centers).

II. The Two Major AI Super PACs: Wealthy and Successful, with Significant Influence

There are currently two major AI super PACs active:

  • "Lead the Future": Raised $125 million (donors include OpenAI co-founders and Anthropic), spending $24 million to support 28 candidates, with 25 winning.
  • "Public First Action": Raised $80 million (with Anthropic contributing $20 million directly), spending $20 million to support 11 candidates, resulting in 10 wins.

These high win rates indicate that either their selection of candidates is accurate or money plays a significant role—after all, candidates need funds for advertising and campaigning, and AI companies' donations provide much-needed support.

III. The Debate Over AI Regulation: Federal Unification vs. State Autonomy

Although both PACs support "moderate regulation" (e.g., protecting minors), they disagree on core issues:

  • "Lead the Future": Advocates for unified national AI regulations. Why? Companies fear that having different rules in each state (for example, California requiring AI transparency while Texas does not) would increase costs due to the need to adapt to various regulations. Unified rules would simplify things.
  • "Public First Action": Prefers states to formulate their own laws. They argue that federal regulations should be comprehensive and reasonable; otherwise, state laws take precedence. They believe that different states have different needs—more stringent regulations in tech-rich states and more relaxed ones in smaller states are more appropriate.

In short, one group wants a national standard, while the other prefers states to make their own choices.

IV. Rare Consensus Between Both Parties: AI Could Disturb Election Fairness

Usually, the two parties have significant disagreements, but on this issue, they agree that AI could potentially interfere with elections.

Members of both parties cite research indicating that over two-thirds of responses from AI chatbots to voter inquiries about candidates are incomplete or incorrect. For instance, when asked about a candidate's stance on AI regulation, a bot might only provide partial information or lie, leading to misleading voters.

Therefore, they are calling on agencies such as the Department of Homeland Security and the Justice Department to collaborate to ensure the accuracy of AI chatbot communications and prevent them from undermining election fairness.

In Conclusion

AI companies are spending money on elections to secure policies that benefit them. The two major super PACs have considerable influence, but they differ on regulatory approaches. Moreover, AI itself could become a source of election interference. All these factors indicate that AI is no longer just a technical issue; it has become a hotly debated and complex topic in American politics.