The Critical AI Report, January 2025

Stargate vs DeepSeek, Top AI privacy risks, AI's impact on critical thinking, and more. Read this or get left behind.

Greetings all,

Welcome to the second edition of The Critical AI Report, a monthly breakdown of the news analysis, academic research, investigative reporting, and key stories shaping critical thought around AI. And what a month it was. I could have comfortably filled this space with news, reports and analysis published in the last week alone, given the rapid-fire news cycle that careened from Stargate to DeepSeek to a market shock that’s rippled across the tech sector—but we take the long view here. And 2025 is off to a predictably chaotic and unsettling start, with plenty of crucial work to parse.

In the AI policy community, there’s been, perhaps unsurprisingly, a lot of thinking about how to approach AI lawmaking and policy under the shadow of a Trump administration. Trump, after all, immediately rescinded Biden’s executive order calling for an AI Bill of Rights, and will likely be uninterested in regulating the American AI industry in any serious way.

Where the Blueprint for an AI Bill of Rights used to be. See: https://www.whitehouse.gov/ostp/ai-bill-of-rights/

To that end, in January, I attended a private gathering of AI policy experts and California lawmakers organized by TechEquity, where I spoke about the history of the Luddites, the politics of big tech, and AI’s impacts on the workforce. It’s clear now that if there *is* going to be any near-term policy that meaningfully protects American consumers and workers from the excesses of the AI industry, it’s going to be enacted on the state level. (The EU has already pursued AI regulation, which is one reason you often hear tech CEOs complaining about Europe.) And there are, despite everything, reasons to be hopeful on this front. A lot of bills are taking shape, aimed at protecting workers’ privacy, shielding consumers from AI-generated misinformation, and more. (And I do mean *a lot*—California legislators write tons of bills, and it was a running joke among legislators just how many AI bills were coming down the pike.)

Alondra Nelson, the who led AI policy in the Biden Administration, and who co-authored the AI Bill of Rights, was in attendance, and rather than despair over Trump’s AI policy wrecking ball, struck a tone of resilient optimism. Executive orders don’t last, she told me, and she knew this was coming. Yet the foundation of their work remains, and even a number of red states are beginning to adopt parts or all of the AI bill of rights, regardless of Trump’s stance.

California, recall, is the fifth largest economy in the world on its own—good laws here can set important precedents and shape global policy. Which is precisely why we need to keep thinking critically about AI, investigate its impacts, and push back when necessary, and possible. To that end! This month’s report will dive into the following:

Stargate vs DeepSeek: Why Silicon Valley will stay the course with a bigger-is-better approach to AI despite shockwaves from a Chinese startup

Five major privacy concerns experts have about generative AI

Does AI inhibit critical thinking? What one major study reveals

The UK is aiming to accelerate AI with pro-industry policies. How artists, creatives, and Paul McCartney are pushing back

AI is using so much power it may be putting your household appliances at risk

Plus, how Los Angeles residents used AI during the wildfires, how Meta’s use of AI bots on Instagram is unsettling users, and a new study that shows people with a lower AI literacy are actually more likely to want to use it.

This report is for paid subscribers. It takes many hours to research, synthesize, and write, and I’ve been told I need to paywall some stuff if I’m going to get serious about this. Thanks as always to everyone who already reads and supports this work; you make it possible. If you haven’t upgraded to paid, and you find value in this work, I’d love to have your support, at the cost of a cup of coffee a month. And if you can’t afford to subscribe and would like access, just shoot me an email, especially if you’re a student, a worker impacted by AI, or an organizer. I’m happy to share. Onwards!