The AI Doc

Audiences need authentic voices now more than ever. Use your craft to show what's at stake, and what's worth protecting. Need help getting started? We provide financial support to content creators aligned with our message.

The "inevitable progress" story serves the companies building AI, not the people affected by it. Interrogate the claims. Make the counterargument impossible to ignore.

AI-generated art, music, and writing are being built on your work without your consent. Advocate for your rights and the rights of every creator who comes after you.

Most people haven't seen how convincing deepfakes, synthetic media, and AI-generated content have become. Shock them into paying attention.

AI risk isn't just a tech story. It's a labor story, a parenting story, a democracy story. People are starting to see how AI is going to affect them—our role is to help them connect the dots. Let's take it beyond the tech bubble.

Dystopian stories are easy. What's harder—and what we desperately need—is compelling, plausible visions of a future where AI genuinely serves people.

We require safety testing for cars, drugs, and aircraft. AI systems that can manipulate, deceive, and displace millions deserve at least the same level of scrutiny, and probably far more.

Big Tech is lobbying hard to stop states from protecting their own residents. Don't let federal frameworks become a ceiling that blocks stronger local safeguards.

AI companions are being deployed to kids with no safety testing, no age verification, and no accountability. Existing child protection laws don't cover this. New ones must.

AI companies are explicit: the goal is to automate human labor. Mandate transparency on job displacement, and ensure workers have a seat at the table.

The AI industry wields enormous financial power. Don't let them convert that into political power so easily. Talk to policy and technical experts who aren't getting their paychecks from Big Tech.

AI risk doesn't stop at borders. Support multilateral efforts to prevent a global race to the bottom on safety. History shows that when the stakes are high enough agreements can be made even among adversaries.

Over 130 countries have called for new rules on autonomous systems. Removing humans from decisions to kill is not a future risk—it's a present one, and the diplomatic groundwork already exists.

Over 130,000+ people have declared their wish for a pause on superintelligence until it can be built safely, and with public buy-in. Other leading AI company staff have already signed, and new signatures carry disproportionate weight. You can even sign anonymously.

1,300+ employees of frontier AI companies have asked the US government to back an international effort to build the technical and governance tools needed to deliberately pace the frontier. It's signed by people at OpenAI, Anthropic, Google DeepMind and Meta AI — this one is written by insiders, for insiders.

Concerns raised individually get managed away. Raised collectively, they move roadmaps. Find your counterparts in other teams and other AI companies.

Whistleblower protections exist, including under California's SB53. Confidential resources include: psst.org, The Signals Network, the AI Whistleblower Initiative. Talk to them before the moment of decision, not during.

Even the heads of AGI companies have expressed that they are nervous about the pace of progress and the risks of superintelligence; encourage them to talk to each other about calling off (or at least pausing) the race for superintelligence. Find out what your company's policy and communications teams are up to; if you don't like it, and your company allows, speak up.

Recursive self-improvement without authorization, autonomous self-replication, shutdown-resistant systems. "I was told to" has a shelf life.

Do you know what commitments your employer has made? Are they keeping up? How are they performing on basic safety and security measures? In some cases, your employer might be falling behind in ways that are straightforward to address.

Alignment, interpretability, corrigibility, and controllability remain unsolved. Fundamentally safe architectures don't yet exist. Formal verification is nascent. The field needs rigorous technical work on making advanced systems safe, not just capable.

Your expertise carries weight that no campaign can replicate. When you see dangerous shortcuts or misleading safety claims, say so loudly and on the record. Join thousands of AI researchers calling for a pause on superintelligence until it can be built safely and with public support.

We're only going to get the kind of AI we want if we can test it and assure users it has the right properties. Without this, governance frameworks have nothing to certify.

If a system's capabilities can't be predicted or its behavior can't be reliably constrained, the responsible choice is not to deploy it. Normalize that standard. Talk about red lines for AI development.

Policymakers need researchers who can translate capabilities and risks into language that informs regulation. Your involvement in the policy process is essential.

Societies, professional organizations, and unions allow you to connect with your fellow researchers and workers, and potentially participate in norm-setting and other collective actions.

Workers losing jobs, families targeted by AI companions, communities flooded with disinformation… these stories need to be at the heart of the policy conversation.

AI safety has remarkable bipartisan support. Partner across political, religious, and cultural divides to demand commonsense regulation that serves everyone.

Scrutinize corporate safety claims. Share the results with your networks and demand that companies meet the standards they advertise.

AI harms fall hardest on those with the least power: children, low-wage workers, marginalized communities. Ensure their voices shape the rules.

Decisions about AI's role in society are being made by a handful of executives. Demand public consultation processes and meaningful citizen input on AI governance.

Nobody will be left unaffected by the change that's coming. Your members deserve to know, and take a stance. Groups of all kinds have taken a position on AI, including the AFL-CIO, Encode Justice, American Federation of Teachers, and more.

AI companies spend enormous sums lobbying against regulation while claiming to support safety. Trace the gap between their public statements and their political spending.

The human cost of AI replacement is already real—in call centers, newsrooms, creative industries, and beyond. These stories are underreported and urgent.

AI companies routinely fall short of their own safety commitments. Our AI Safety Index and other trackers provide data-driven accountability reporting you can build on.

Algorithms are already deciding who lives and who dies on the battlefield. This is one of the most consequential and least-covered technology stories of our time.

AI-generated journalism erodes public trust and your profession. Report on the threat, and advocate for standards that protect journalistic integrity and employment.

AI is a complex topic, but it doesn't take years or even months to get up-to-speed. If you're learning about this for the first time, this website will give you a great start.

If your AI products are controllable, transparent, and respect user data, that's a competitive advantage. Market it. Customers and regulators are paying attention.

Commonsense safety standards create a level playing field and protect responsible companies from being undercut by reckless competitors racing to deploy unchecked systems.

You can build powerful, profitable AI tools without pursuing uncontrollable general intelligence. The most valuable AI solves specific problems safely and reliably.

Deploy AI to augment your employees, not replace them. Companies that treat AI as a tool for empowerment, not elimination, will attract and retain the best talent.

If you rely on AI systems from the leading AI companies, demand to know how they're tested, what risks they carry, and what happens when they fail. Your liability is real.

A growing movement of companies, organizations, and public figures are backing pro-human principles for AI development. Add your voice to the business case for safety.

Frontier AI systems built by a handful of companies carry risks their creators can't fully predict. Invest in the technical expertise and institutions needed to independently evaluate and monitor AI systems deployed within your borders.

International AI governance is being defined now. You have the diplomatic credibility and the incentive to champion safety standards that serve broader interests, not just Big Tech's. Build shared frameworks for safety, evaluation and governance that serve collective rather than competitive interests.

Unchecked dependence on a few foreign AI providers creates strategic vulnerability. Pursue policies that maintain meaningful control over the technologies shaping your economy.

The case for prohibiting lethal autonomous weapons is overwhelming, and your leadership could be decisive.

Collective bargaining power matters. Form alliances that survive political pressure and hold interests in tension with like-minded states to set shared safety requirements and actually influence advanced models' training and deployment.

Join forces to build native AI systems that follow the pro-human path and solve concrete problems defined by your own national priorities.