In November 2022, OpenAI released ChatGPT and ignited an AI arms race. Google, Meta, Anthropic, and a Chinese startup called DeepSeek all raced to compete. In barely three years, AI has gone from a novelty that could write passable college essays to a technology that can generate lifelike video, clone human voices, and write software code so effectively that people with no programming experience are building apps on their own. It is a revolution, and most people have no idea how fast it is moving.

And the stakes are getting higher. According to reports, the Pentagon told Anthropic to remove the safety guardrails from its models — including restrictions on domestic mass surveillance and autonomous weapons — or lose its government contract. Anthropic refused.

Defense Secretary Pete Hegseth labeled Anthropic a “supply chain risk to national security,” and President Trump ordered every federal agency to stop using its products. Hours later, OpenAI swooped in to secure its own Pentagon deal, agreeing to terms that critics say offer far weaker protections.

The threat is not only coming from the government. A Stanford University study found that, at the time of the study, six major AI companies all used consumer chat conversations to train their models by default. Hundreds of millions are handing over intimate information governed by privacy policies they will never read. But most have no idea this is happening — even as they sense that something is going wrong.

A Pew Research survey found that 57% of Americans rate the societal risks of AI as high. Two-thirds believe it will eliminate more jobs than it creates. And 60% simply do not trust AI at all. These fears are not wrong. But fear is not the same as understanding.

Most people cannot explain what a large language model does or what happens to the data they share with a chatbot. They know enough to be alarmed. They do not know enough to protect themselves. AI is creating a new and more dangerous digital divide — not between those who have access to technology and those who do not, but between those who understand it and those who are subject to it.

Between a federal government demanding AI without safeguards and corporations collecting vast amounts of personal data with little meaningful transparency, we face a growing gap between how AI is being deployed and how well citizens understand it. Comprehension is not just an educational goal — it is the precondition for citizens to demand better. An uninformed citizenry doesn’t know what it doesn’t know.

New York is well positioned to respond. Gov. Hochul has signed the RAISE Act into law, joining California as the only states establishing transparency requirements for frontier AI developers. Her Empire AI initiative is aiming to build one of the most powerful academic AI supercomputers in the nation. And she just announced a new commission to study the effects of AI on the economy.

In the Legislature the proposed Artificial Intelligence Literacy Act would fund AI education in public schools, community colleges, and nonprofits statewide. These are important steps. But a program that reaches classrooms over the coming years will not help the millions of New Yorkers worried right now about whether AI will take their jobs, flood elections with misinformation, or be used in ways that affect their civil liberties without their knowledge.

The answer is not to stop the technology. It is to ensure that every New Yorker has enough understanding to engage with it on their own terms.

There is a proven model for this: in 2018, Finland published a free online course called “The Elements of AI,” designed to teach 1% of its population the basics of artificial intelligence. It was so successful that Finland offered it to every EU member state, ultimately enrolling more than two million students from 170 countries. Crucially, it was built by a public university with public funding. What we need is a publicly funded American version, available to everyone.

Hochul has rightly begun to treat AI as a public safety imperative. She should insist that the AI Literacy Act be expanded to include a mass public education campaign — let’s call it AI Education for All — one that offers a free online course modeled on the Finnish approach, promoted through state government websites, public media, social media, MTA stations, buses, trains, and public libraries.

The governor could start tomorrow by directing the Division of Homeland Security and Emergency Services — whose mandate already covers human-caused threats to New Yorkers — to make AI literacy part of its public safety mission.

In New York City, Mayor Mamdani is uniquely positioned to amplify that effort. No mayor in America has demonstrated a greater command of social media as a tool for civic engagement and government communication. He amassed more than a million followers on Instagram and 3.5 million on TikTok during his campaign, and as mayor is already using social media, LinkNYC kiosks, and content creators to communicate directly with New Yorkers.

Mamdani could direct his Office of Technology and Innovation — whose mandate includes digital equity — to make AI literacy a citywide priority, and ask his Office of Mass Engagement to amplify the message. During the COVID pandemic, government at every level mobilized to educate the public. The same urgency is needed now.

Some will argue that regulation alone is the answer. We certainly need new laws protecting citizens from AI-assisted government and corporate surveillance. But regulation is not a substitute for comprehension.

Artificial intelligence is advancing too fast for any legislature or regulatory agency to keep up. The only durable safeguard is an informed citizenry. Finland grasped this eight years ago.

Hochul and Mamdani have the tools and the skills to help protect all New Yorkers and create a model for other states to adopt to protect the nation. It is time that they use them to give citizens the knowledge and power they need to protect themselves and their futures.

Rasiej is the founder of Civic Hall @ Union Square and serves on the New York City Office of Technology and Innovation’s AI Advisory Board.