Blog

The companies building the most powerful artificial intelligence systems are making an unusual request: they want governments to set rules for their industry. For executives accustomed to watching technology companies resist oversight, the appeal sounds almost contradictory. Why would a business racing for customers, capital and market share invite regulators into the room?

There are credible safety reasons. There are also commercial reasons to ask for clear rules, and reasons for investors and competitors to scrutinize the fine print. The central question is no longer whether AI needs oversight. It is who writes the rules, which systems they cover and whether smaller companies can afford to comply.

The Risks Have Moved Beyond the Laboratory

AI can now generate convincing text, voices and images at scale. It can assist with hiring, lending and other decisions that affect people’s opportunities. Newer systems can also carry out multistep tasks with less human direction. A mistake or misuse can therefore spread far beyond a single user.

The record of reported problems is growing. Stanford University’s 2026 AI Index counted 362 documented AI incidents in 2025, up from 233 in 2024. That count reflects recorded incidents, not every harmful outcome, but it shows why calls for testing and disclosure have gained force. Regulators are also weighing discrimination, fraud, privacy, copyright and potential cybersecurity or biological risks. Claims that future AI could pose a catastrophic threat remain contested; harms such as deceptive content and flawed automated decisions are already concrete concerns.

The money involved raises the stakes. Stanford estimates that global corporate AI investment reached $581.7 billion in 2025, an increase of 130% from the previous year. U.S. private AI investment totaled $285.9 billion, and 1,953 newly funded AI companies emerged in the United States. When investment moves at that speed, businesses have incentives to release products quickly, while customers and governments have reason to ask how those products were tested.

Why the Builders Want Rules

Some AI leaders say the technology’s capabilities demand external oversight. In 2023, OpenAI leaders proposed international supervision for future systems far more capable than today’s models. That position can be taken seriously without assuming every proposed rule is well designed.

Regulation can also solve problems for the companies requesting it. A common framework could give enterprise customers a clearer basis for evaluating vendors. It could replace a patchwork of requirements with more predictable obligations. Shared testing and reporting standards could help companies demonstrate how they manage risk. Following a safety rule, however, would not automatically shield a company from liability if its product caused harm.

The commercial benefit depends on the details. A rule requiring developers to document serious incidents might improve accountability across the market. A costly licensing regime could have a different effect if only a few firms have the computing resources, legal teams and capital to meet its conditions. Public concern and business self-interest can exist at the same time.

Why Wall Street Is Skeptical

Investors have learned to examine what a company gains when it asks for regulation of its own product. If an AI developer says its systems are sufficiently powerful to require special oversight, investors may hear two messages: the technology has enormous commercial potential, and its risks could create substantial legal, operating and reputational costs.

There is also a competition question. In a 2025 study of major cloud providers’ partnerships with AI developers, the Federal Trade Commission identified arrangements that could increase developers’ switching costs and affect access to resources needed to compete. Against that backdrop, investors have reason to ask whether a proposed safety standard protects customers, reinforces the advantages of established firms, or does both.

A further theory is that regulation could make a crowded market easier for its leaders to defend. If every new entrant must pay for extensive audits, specialized staff and computing intensive tests, a well funded incumbent may absorb those costs more easily than a startup. That is a plausible competitive effect, not proof that any particular founder is seeking to exclude rivals. Investors should judge a proposal by its thresholds, costs and enforcement provisions, rather than by a company’s stated intentions alone.

A Break With the Tech Playbook, With an Important Qualification

The appeal for oversight contrasts with the familiar Silicon Valley approach of launching rapidly and arguing over rules afterward. But the distinction should not be overstated: businesses have long supported regulations that create certainty or raise the cost of entry for competitors. What makes AI striking is how openly some of its builders have warned about risks from the technology they are selling.

Even within the industry, there is no single position. Developers disagree over licensing, model disclosure and whether rules aimed at the most capable systems could hinder smaller or openly available models. The sharpest disagreement is often about where a requirement begins: at a system’s computing scale, its demonstrated capabilities or its use in a consequential setting such as employment or credit.

The legal landscape is equally divided. The EU AI Act entered into force in 2024, with its main provisions becoming applicable in August 2026 and other obligations following different schedules. The United States still has no single comprehensive federal AI law. State measures and existing laws governing areas such as consumer protection and employment remain part of the picture. The U.S. executive order that had required certain developers to share safety test results with the federal government was revoked in January 2025; describing it as a current nationwide requirement would be inaccurate.

What This Means for U.S. Hispanics

For Hispanic Americans, the debate has immediate economic consequences. Bureau of Labor Statistics data for August 2026 put the Hispanic or Latino labor force at approximately 35.7 million people, with nearly 34 million employed. Rules governing AI assisted hiring, workplace management and access to training will therefore affect a substantial share of the U.S. workforce.

The issue extends to business ownership. An analysis of Census data by Brookings counted 465,202 Latino or Hispanic owned employer businesses in 2022, representing 7.9% of all U.S. employer businesses. Affordable AI tools can help these firms translate materials, serve customers, analyze operations and compete with larger companies. If compliance costs are passed along through higher prices or restricted access, smaller businesses could face a narrower set of choices.

Fairness also requires attention to how systems perform for the people using them. A hiring or lending tool should be tested for discriminatory outcomes; a customer service system should work reliably in Spanish and across dialects. Those needs do not point to a single preferred regulation. They do show why Hispanic workers, entrepreneurs and consumers have a stake in both effective safeguards and open competition.

AI developers should be required to provide lawmakers with technical evidence about their systems, including testing methods, failures and known risks. But workers, consumers, independent researchers and smaller competitors must have an equally meaningful voice. The test of any regulation is whether it reduces harm without allowing the largest companies to write barriers that protect their market position.

Sources

• Brookings Institution. (2025). Charting the surge in Latino or Hispanic-owned employer businesses.
• European Commission. (2026). AI Act: Shaping Europe’s digital future.
• Federal Trade Commission. (2025). Partnerships between cloud service providers and AI developers.
• OpenAI. (2023). Governance of superintelligence.
• Stanford Institute for Human-Centered Artificial Intelligence. (2026). The 2026 AI Index report.
• U.S. Bureau of Labor Statistics. (2026). The employment situation: August 2026, Table A-3: Employment status of the Hispanic or Latino population by sex and age.
• The White House. (2025). Initial rescissions of harmful executive orders and actions.