Red Tape or Red Threat?

Right now, the world is at a crossroads: Do we supercharge AI innovation, or do we clamp down before it’s too late? The debate is taking place, live, on two opposite paths. One is taken by Europe, the other by China.

If we shackle ourselves Europe-style, China will win. But what does it really mean if China wins this race?

Picture this: Our children and grandchildren growing up in a world where hidden algorithms watch, judge, and determine their fates. Imagine the foundations of their lives - their schooling, their bank accounts, even their social circles - resting on platforms built on surveillance. A teenager’s shot at college, a job, a future, would hinge on a “reputation score” that rewards compliance and harshly punishes dissent. Criticize corruption? Support the wrong cause? Associate with the wrong friend? The algorithm quietly rations the opportunities they’ll have in life.

China’s ambitions are no secret. It has made global leadership in AI a core national priority and set explicit milestones paired to that goal. As the foundational document of their strategy puts it:

By 2030, China’s AI theories, technologies, and applications should achieve world-leading levels, making China the world’s primary AI innovation center.” (Source: State Council of the People’s Republic of China, “New Generation Artificial Intelligence Development Plan,” 2017)

But here’s the urgent truth: The choice before us is not “become Europe or become China.” There is a third way, a way that defends both innovation and liberty.

The sane, bold middle path is simple: regulate AI’s dangerous uses, not the whole frontier. Don’t bury clever new ideas under a ton of paperwork. Focus laser-like scrutiny where it counts, on mass surveillance, automated repression, election manipulation, biological threats, and AIs that run critical infrastructure. These are where we need bans, firewalls, and heavy licensing. But let the vast majority of AI, such as tools for work, learning, or creativity, flourish, protected by straightforward rules for safety, privacy, and transparency.

We can weave strong safety rails into the most advanced models, without slamming the brakes on progress. However, above a certain level of AI impact, we can require companies to do stress tests and build in reliable “off switches.” In high-stakes fields like healthcare, banking, or policing, there is a demand for airtight protection from mistakes. Crack down, hard, on deepfakes and outright manipulation. There’s so much at stake that long prison sentences would be given, maybe even a lifetime sentence as a deterrence, sentences that demonstrate that this behavior is never tolerated.

The real answer isn’t just to manage risks better. It’s to outpace them. That means staking serious long-term bets on homegrown AI innovation. Boost research, attract the planet’s best talent, and set the standard with public-interest AI, including tools, transparency, and ethics that lift the whole industry. Democracies win by pairing breakneck speed with high trust.

We develop our AI, using national standards as opposed to piecemeal and likely ineffective state standards, and then we provide access to it for all governments that value the well-being of their citizens: no social credit scores, no ethnic profiling, no turnkey surveillance states. Access to our markets and cloud goes only to those who share those rules. Our goal is a clarion summons to make the AI we, as a nation, develop be the best, safest, most irresistible option anywhere.

If we choose this path, we don’t surrender the future to China, and we don’t sabotage ourselves with paralyzing rules. We strike fast, aim smart, and have guardrails that aren’t cages. If we get this right, the next generation can invent, dissent, and dream on their own terms, never at the mercy of a Chinese social credit score.

Mitzi Perdue is a fellow at the Institute of World Politics and the Co-Founder of Mental Help Global, a philanthropy that uses AI to support mental health.