America's AI action plan isn't about ethics, it's about empire

The Trump administration has officially dropped it's AI action plan. On paper, it reads like a strategy for technological innovation. But if you scratch the surface, you'll find something more unsettling: a blueprint for dominance, not democracy.

This isn't just an investment in smarter machines. It's a full-throttle political play to win the global AI race economically, militarily, and ideologically. And the stakes are enormous. Because while the plan talks about "freedom" and "innovation" what it quietly sidelines are the safeguards that protect people: regulation, equity, accountability, and truth.

The real goal? Outpace China at any cost

The language of the action plan is steeped in competition. America must "win the AI race." No surprises there. The rhetoric echoes Cold-War-era nationalism: think space race, but for algorithms. And in this race, ethics are framed as obstacles, not essentials.

The plan scraps previous commitments made under the Biden Administration, including hard-won rules around transparency, safety testing, and harm reduction. It even threatens to punish states that try to pass their own stronger AI regulations, potentially cutting off funding if they don't fall in line with the federal vision. That's not innovation, it's coercion.

Idealogical cleansing hidden behind "objectivity"

One of the most disturbing aspects of the plan is its quiet rewriting of what counts as "truth." Federal AI contracts will now only go to systems that adhere to a vague standard of "objective truth." Sound harmless, until you realise who gets to decide what the truth is.

In practice, we're already seeing key issues like disinformation, climate science, and racial justice disappear from offical AI frameworks. This isn't accidental, it's strategic. It means AI systems developed under this plan will likely ignore or erase complex social issues in favour of political convenience.

Open source, but closed accountability

The plan promotes open-source AI as a pillar of innovation. And yes, open access can be powerful for collaboration and community-led growth. But in this context, "open" doesn't mean safe. These models can be deployed at scale with minimal oversight, no requirements for harm mitigation, and no investment in long-term safety.

This is open source in the same way a demolition site without warning signs is "accessible." Without checks and balances, openness becomes a liability not a virtue.

Who benefits? Not workers. Not the public.

The action plan frames itself as worker-first. But the definition of "worker" is narrow focused on those who build AI infrastructure (engineers, developers, chip manufacturers), not the broader labour force most vulnerable to automation.

There's little acknowledgement of the call centre agents, logistics staff, or knowledge workers whose jobs are being transformed or erased by these same technologies. Their voices are entirely absent. Once again, those closest to the impact are furthest from the table.

Exporting power, not partnership

The US isn't just aiming to dominate AI domestically. It's laying the groundwork to export it's entire AI stack to "trusted allies".

We've seen this pattern before: military equipment, surveillance tech, cybersecurity tools, all exported in ways that deepen dependency and reduce local sovereignty. AI is simply the next layer of techno-nationalism. For countries like Aotearoa New Zealand, this is a warning shot. If we're not building local capacity, we're buying into someone else's values and power structures.

So where does that leave us?

The plan isn't just policy, it's precedent. And while it's easy to dismiss it as just another American move, its influence will ripple far beyond US borders. It sets the tone for how the global AI ecosystem develops.

And if you're someone who believes AI should be transparent, democratic, and just, now is the time to start paying attention. Because plans like this don't stay contained. They spread through trade deals, tech partnerships, and policy harmonisation.

What can we do?

• Stay informed. Read the reports, not just the headlines.

• Talk about it. These aren't abstract debates, they affect your data, your job, your kids' schools.

• Push for local leadership. New Zealand can (and should) take a different path. But we need to be proactive, not reactive.

• Support ethical innovation. The future doesn't need to be a race. It can be collaboration.

The bottom line? The Trump AI plan isn't about responsible innovation. It's about winning, regardless of who loses. That should be a worry to us all.

What kind of AI future do you want to be part of?