We need a more balanced approach to developing and deploying AI

Eric Schmidt and Selina Xu, We Must Address the Growing Rage Against the A.I. Machine, NYTimes, July 11, 2026.

On the whole US AI developers want to move as fast as possible while the public is increasingly skeptical and opposed. The Chinese are more cautious and want to develop AI while preserving social stability thus the Chinese public are more optimistic about AI than Americans.

Dismissing A.I. entirely would be a mistake. The technology can help diagnose diseases, predict protein folding, improve farming, forecast disasters better, design new materials, accelerate scientific and drug discovery and power robots in dangerous environments to improve human safety (such as in space, firefighting and minefields).

Yet how technology spreads is never inevitable. If A.I. is viewed as benefiting the few at the expense of the majority, then the public will rage against the machine. And A.I. won’t be able to make our lives better in the long run if it cannot survive in the short term. The real challenge, then, isn’t whether the United States or China will build an overwhelming, insurmountable advantage over the other. It’s whether either can figure out how to realize the benefits of A.I. without ripping apart its social fabric. Neither has found the answer yet.

Silicon Valley tech executives and policymakers across the country are waking up to that fact. States have introduced dozens of bills this year to put safety and privacy guardrails around A.I. The Trump administration issued a new executive order that seeks to give the government more oversight over new models before they’re released to the public. [...]

To get to the other end of the tightrope, we need radical and incremental solutions alike. Here’s one to start: a populist A.I. agenda that treats the technology as a public project. Just as NASA made space a national mission rather than a private one, the government should now do the same for A.I. to ensure that its benefits reach the public, not just the companies building it. Here are some concrete ways to do that:

One, treat some A.I. profits as a shared resource and distribute them directly to citizens. This can be in the form of a sovereign wealth fund, seeded by contributions from A.I. companies in the form of stock or cash. [...]

Two, support public-interest A.I. models that the private sector might not be incentivized to build. Examples include models to help citizens navigate government services and benefits, to help them gain legal aid or to help educate their children. We strongly believe in the importance of open-source A.I., which anyone can freely use and modify. [...]

Three, regulate one area that attracts bipartisan agreement: the way children and teens interact with A.I.

There's more at the link.