Tax the Bots, Save the Humans: OpenAI Drops Policy Wishlist for the Intelligence Age

ChatGPT creator OpenAI is urging world leaders to get ahead of the curve before advanced AI reshapes everything. In a paper titled "Industrial Policy for the Intelligence Age: Ideas to Keep People First," released Monday, the company argues that rapid AI advances could fundamentally restructure economies—and that tax systems, labor policies, and social safety nets need to evolve accordingly. Basically, they're asking governments to start sweating now about the robot apocalypse before it shows up at the unemployment office.

"No one knows exactly how this transition will unfold," OpenAI noted. "We believe we should navigate it through a democratic process that gives people real power to shape the AI future they want, while preparing for a range of possible outcomes and building the capacity to adapt." Translation: we're all flying the plane while building it, and nobody's sure if it has wings.

The company contends AI could boost productivity and accelerate scientific discovery, but also warns of potential labor market disruption and wealth concentration if policies don't keep pace. Governments should start prepping now for shifts in work, income, and economic growth, the document says. Nothing says "fun policy debate" like explaining to voters why their kids need to retrain as prompt engineers at age 45.

Key policy ideas include treating AI access as a foundational economic resource—comparable to global literacy initiatives—modernizing tax systems for automation, and creating mechanisms for citizens to share in AI-driven economic gains. Imagine universal basic income, but instead of printing money, you're printing value from machine learning models. Fun stuff.

"The promise of advanced AI is not just technological progress, but a higher quality of life for all," OpenAI wrote. "Living standards should rise, and people should see material improvements through lower costs, better health and education, and more security and opportunity." Theoretically, anyway. Historically, technology tends to make the rich richer and everyone else pretty good at doomscrolling.

The paper also calls for stronger worker protections and expanded social support if tech changes trigger sudden job losses, plus oversight tools like frontier model auditing, incident reporting systems, and "model-containment playbooks" for dangerous AI scenarios. Yes, there's apparently a playbook for when the AI decides to go full Skynet. Someone's been watching too many movies.

"If AI winds up controlled by, and benefiting only a few, while most people lack agency and access to AI-driven opportunity, we will have failed to deliver on its promise." A bold statement from the company that literally built a paywall around intelligence.

This policy push lands amid turbulence for OpenAI CEO Sam Altman, who's facing renewed scrutiny after a New Yorker investigation revealed that co-founder and then-chief scientist Ilya Sutskever wrote internal memos in 2023 accusing Altman of being deceptive about safety protocols and key operations. The board subsequently fired Altman, concluding he hadn't been "consistently candid" with them. Nothing like a little corporate governance drama to spice up your Monday morning.

The firing sparked internal chaos—employees threatened to walk, and investors like Josh Kushner threatened to withhold funding until Altman was reinstated. The report highlighted deep divisions over governance and safety, with insiders including Sutskever and Anthropic co-founder Dario Amodei arguing Altman prioritized growth over the company's original safety-focused mission. OpenAI didn't respond to Decrypt's request for comment. Shocking.

Meanwhile, Anthropic has filed paperwork with the Federal Election Commission to create a political action committee, signaling a deeper push into U.S. politics as AI policy debates heat up. The San Francisco-based company registered the Anthropic PBC Political Action Committee, AnthroPAC, in a Friday filing. The committee is structured as a separate segregated fund tied to the company. Nothing says "we care about humanity" like forming a PAC to lobby the people deciding humanity's future.

In separate news, Anthropic researchers say they've identified internal patterns in one of their AI models that resemble representations of human emotions and influence system behavior. In a paper, "Emotion concepts and their function in a large language model," published Thursday, the company's interpretability team analyzed Claude Sonnet 4.5 and found clusters of neural activity tied to emotional concepts like happiness, fear, anger, and more. So the AI might be having feelings. We're definitely not overthinking this one.

On the quantum front: quantum computers can't break Bitcoin's cryptography today, but new advances suggest the gap is closing faster than expected. Progress toward fault-tolerant systems raises stakes for "Q-Day"—the moment a sufficiently powerful machine could crack older Bitcoin addresses and expose over $711 billion in vulnerable wallets. Long seen as a distant threat, Q-Day snapped into sharp focus in March 2026, with multiple research papers suggesting quantum capabilities may advance quicker than anticipated. Sleep well, crypto holders.