What Trump Gets Right About Ai Risks While Fighting Global Rules

What Trump Gets Right About Ai Risks While Fighting Global Rules

Donald Trump just told world leaders at the APEC summit in Shenzhen that artificial intelligence is "almost completely good," yet warned that nations have to watch it closely and manage its "bad elements."

Saying that out loud on Chinese soil—right before sitting down for high-stakes talks with Xi Jinping and Vladimir Putin—is a pretty bold balancing act. Honestly, it captures the central headache of AI policy right now. You want the economic rocket fuel. You do not want the blowback. And if you push too hard on international rules, you might hand your biggest rival a free pass.

Look closely at what Trump actually said in his plenary remarks. He called AI—or "superintelligence," his preferred label since a September UN address—the biggest technological shift since the Industrial Revolution. At the same time, he pushed back hard against what he called "globalist regulation" that could control or stifle growth.

That tension isn't going away anytime soon.

Why Managing Bad AI Elements Matters More Than Global Treaties

Most headlines treat AI regulation as an all-or-nothing fight between free-wheeling innovation and heavy-handed bureaucracy. That framing misses what is actually happening on the ground.

When Trump warns about "bad elements," he's pointing to real operational and security headaches without signing onto a sweeping UN-style rulebook. Look at the week surrounding the Shenzhen gathering. Washington moved to curb remote access to advanced American AI chips and servers for Chinese firms. Just days earlier, on November 15, the United States and nine partner nations—including India—signed the San Francisco Declaration on the Secure Foundations of Enduring Prosperity during the Pax Silica Founders Retreat, focusing on supply chain security and critical infrastructure for emerging superintelligence systems.

Meanwhile, real-world friction is piling up fast. Reports out this week from the Monetary Authority of Singapore flag how fast-moving AI agents can accelerate financial risks, while security watchdogs and homeland officials keep raising alarms about jailbreaks, grid vulnerabilities, and software upgrades triggering messy outages.

💡 You might also like: what is a computer icon

Here's what I've seen watching tech policy play out over the years: grand international treaties sound great in speeches, and they usually move at glacial speed while the underlying tech changes every six weeks. Targeted controls around hardware supply chains, critical infrastructure, and abusive deployment tend to be way more practical than a single global rulebook enforced unevenly by governments with wildly different interests.

Still, there's a catch with the "light touch on rules, heavy hand on access" approach. Critics argue that if major powers reject shared guardrails, everyone races to the bottom on safety testing just to stay competitive. Even within the U.S., voters have been pushing back on local impacts—like power grid strain and water use from massive data centers—which showed up in recent midterm frustration.

The Superintelligence Label and the U.S.-China Race

Trump's insistence on calling the technology "superintelligence" rather than artificial intelligence isn't just a quirk. It's a framing choice.

In his view, calling it "artificial" makes it sound fake, whereas "superintelligence" signals raw capability and strategic leverage. Whether AI engineers agree with that terminology or not—most researchers draw a sharp distinction between today's narrow and agentic models and true general or superhuman intelligence—the political point is crystal clear. Washington wants to own the narrative of technological leadership.

That race is front and center in Shenzhen. Xi Jinping used the APEC stage to pitch regional cooperation, integration, and shared development across the Asia-Pacific. Trump countered by pitching American economic strength, energy exports—including an approved Alaska liquefied natural gas pipeline aimed at Asian buyers and oil activity in Venezuela—and U.S. dominance in AI.

You can't separate AI from energy and compute anymore. Training and running frontier models takes massive electricity and specialized chips. By tying AI leadership to American energy and chip supply chains, the White House is treating compute and power as strategic assets just like oil or steel.

One mistake people make when reading summit speeches is assuming diplomatic talk equals a settled truce. Meetings with Xi and Putin can open channels on trade, security, or nuclear arms control—where Trump noted how tough it is to get Beijing to curb its arsenal without massive U.S. reductions—while fierce tech competition keeps humming underneath.

Where Innovation Meets Real-World Risk

So if you're running a company, building software, or trying to make sense of where this goes next, what does this split message actually mean?

It means you shouldn't count on Washington or global summits to hand you a neat, settled compliance checklist tomorrow. Rules are being shaped piecemeal through export controls, national security reviews, sector-specific oversight, and court or state-level pushback.

  • Don't treat "no global treaty" as a free-for-all. Even when leaders oppose broad international regulation, enforcement around sensitive hardware, defense uses, fraud, and critical infrastructure can tighten fast.
  • Watch compute and energy bottlenecks. AI scaling isn't just an algorithm story; chip access, server availability, and power costs decide who actually ships and who gets stuck waiting.
  • Test for failure before regulators force you to. Outages, prompt attacks, and runaway automated workflows hurt customer trust long before a new law passes.

If you want to stay ahead of this shift, start with three concrete moves this month:

  1. Audit your AI supply chain and dependencies. Map out which model providers, cloud regions, and hardware vendors your stack relies on, and identify single points of failure if trade or access rules shift.
  2. Put human sign-offs on high-stakes actions. Keep automated AI agents out of irreversible financial moves, production infrastructure changes, or unverified external communications without clear guardrails.
  3. Track both policy and infrastructure signals. Follow chip export rules, data center power debates in your region, and central bank or regulator warnings—those practical pressures will hit your operations long before abstract debates over "superintelligence" get resolved.
VJ

Victoria Jackson

Victoria Jackson is a prolific writer and researcher with expertise in digital media, emerging technologies, and social trends shaping the modern world.