Amid a murky AI regulatory landscape, OpenAI is making moves.

On Wednesday, the AI giant announced its endorsement of two AI safety bills, the Kids Online Safety Act and SB 315, a bill on frontier AI. In its global affairs blog, the Prompt, the company said that the former aims to create stronger protections for young AI users, while the latter supports a "national framework for frontier AI safety."

Though these bills target different aspects of safety, both aim to strengthen safety requirements and accountability.

• The Kids Online Safety Act, or KOSA, sets clear protections for minors' social media use. OpenAI noted that new models necessitate "AI-specific rules," and that the bill is complementary to the work it’s doing at the federal and state level. In the blog, Chief Global Affairs Officer Chris Lehane said, "We can’t repeat the mistakes made during the rise of social media."
• SB 315 takes aim at frontier AI models, calling for clear requirements around safety, transparency, incident reporting and accountability for bleeding-edge systems. The Illinois bill follows similar legislation in New York and California, and OpenAI is supporting the bill to advance "a risk-based approach focused on the most capable models and highest-consequence harms."

Beyond its support for these bills, OpenAI also announced the opening of its Washington, DC AI workshop, aiming to provide a space to foster "much-needed conversations about who will really benefit from AI." The workshop will host training and discussions for elected officials, regulators, civil servants, educators, workers, nonprofits, and industry and community leaders to engage with and discuss the technology.

In the blog post, OpenAI said that, rather than being a commodity, intelligence should become a "global utility," akin to electricity, available to people, businesses and institutions "as much as they need, where and when they need it."

"Our job is to make that intelligence cheaper, better, and more abundant over time – not to dominate every vertical ourselves – so that more people can build, solve problems, and expand what they’re able to do," the blog post reads.

Our Deeper View

AI’s regulatory path is far from set in stone. OpenAI has the ability to impact how that path is shaped by supporting bills early and hosting conversations. While the company says in its post that the aim is not to "dominate every vertical" by becoming a utility, it is undeniable that the company holds an extraordinary amount of power to shape everything from the laws that govern it to the way people use it. And given that the goal of any company in a capitalist system is to increase their bottom line, it’s always important to consider the incentives. OpenAI is especially good at striking the right notes in its rhetoric. We still need to see evidence from OpenAI, and the other leading AI labs, that they have the will to make tough decisions that will support high-minded goals such as "ensuring AGI benefits all of humanity" even when it goes against their economic interests. That's going to become even more challenging when OpenAI, Anthropic, and others become public companies and have to answer to shareholders.