The AI Paradox: From Demanding Deregulation to Begging for Rules

What happened?

According to a Politico report picked up by The Next Web, executives from leading AI companies that generously donated to Donald Trump's campaign — with the promise that he would leave the technology unregulated — are now formally requesting the creation of clear rules. They consider the administration's ad hoc approach to overseeing models more harmful than anything Biden would have done. Among the companies lobbying for formal regulation are OpenAI, Google DeepMind, and Anthropic, which collectively donated over $10 million to pro-Trump political action committees in 2024. The request was formalized in a letter sent to the White House in May 2026, signed by more than 30 industry executives, calling for a risk-based regulatory framework similar to the EU's.

Historical context

During the Biden administration, the Office of Science and Technology Policy (OSTP) issued an executive order on AI in October 2023 that established principles for safe and responsible development, including transparency requirements and safety testing for frontier models. The industry criticized it as excessive, arguing it would stifle innovation. Trump, for his part, promised to repeal it and in January 2025 signed an executive order eliminating Biden's guidelines, opting for a 'hands-off' approach. Subsequently, in March 2025, the Trump administration issued non-binding guidance leaving self-regulation to companies. Now, without a clear framework, companies face legal uncertainty and a lack of guidelines for self-regulation. This shift in stance recalls what happened with social media in 2018, when Facebook and Twitter went from opposing content regulation to requesting it amid public pressure.

Why is this important?

This turnaround reveals a fundamental paradox: the AI industry needs rules to operate with confidence. Without a regulatory framework, investors hesitate, developers don't know what limits to comply with, and security risks multiply. According to CB Insights data, investment in AI startups in the U.S. fell 18% in the first quarter of 2026 compared to the same period in 2025, partly attributed to regulatory uncertainty. Moreover, the absence of rules hinders international cooperation: the EU, with its AI Act (approved in 2024 and in effect since 2025), requires high-risk models to meet strict requirements, and companies like OpenAI have had to delay launches in Europe due to lack of clarity on which regulations apply. In contrast, China has implemented its own regulatory framework since 2023, requiring pre-approval for generative AI models, which has allowed giants like Baidu and Alibaba to launch products with legal certainty.

Consequences for the sector

• Investment slowdown: Regulatory uncertainty discourages venture capital, which needs predictability. According to a Stanford HAI report, AI investment in the U.S. grew only 4% in 2025, down from 22% in 2024, and a contraction is expected in 2026 if clear rules are not established.
• Security risks: Without guidelines, models can be released without adequate controls. In February 2026, a language model from an unidentified startup generated hate speech at scale, prompting a congressional investigation. Experts like Yoshua Bengio have warned that lack of regulation increases the danger of malicious uses, such as automated disinformation or autonomous weapons development.
• Advantage for foreign competitors: While the U.S. lacks rules, China and the EU advance with their own frameworks. The EU has already certified more than 50 models as 'compliant' under its AI Act, giving them preferential access to the European market. China, meanwhile, has approved over 40 generative AI models since 2023, and companies like SenseTime have seen revenues grow 30% thanks to regulatory clarity.
• Legal and compliance costs: Companies face lawsuits for AI-related damages without a clear liability framework. For example, in April 2026, a class-action lawsuit against an AI company for copyright infringement was dismissed due to lack of standards, creating legal uncertainty.

What readers should know

This case illustrates that self-regulation is not enough. The industry, which once saw regulation as an obstacle, now understands it as a market enabler. Readers should watch for upcoming moves by Congress and the Trump administration. In June 2026, the Senate is expected to debate the 'AI Regulatory Framework Act,' a bipartisan bill that could establish a federal AI agency similar to the FDA. Also relevant is California's initiative, which in 2025 passed the 'California AI Safety Act,' requiring safety testing for models trained in the state, potentially becoming a de facto standard. The key lesson is that regulatory uncertainty is more costly than regulation itself, and the pendulum has swung: the same industry that asked for deregulation now begs for rules.

'The pendulum has swung: the same industry that asked for deregulation now begs for rules. The lesson is that chaos is not profitable.' — TheVortiq