Picture this: you answer your phone and hear a familiar voice telling you not to vote. Or you scroll through social media and see a video of a political leader saying something utterly outrageous, something that could change your mind—except the voice is synthetic, and the face has been digitally stitched onto someone else’s body. This is no longer science fiction. As artificial intelligence deepfake technology becomes more realistic and more Americans rely on AI chatbots for everyday information, the risk of being deceived has never been higher. The same tools that can make a singer appear to perform a song they never recorded are now being used in political campaigns, often with the goal of suppressing votes. With the midterm elections only weeks away, New York Attorney General Letitia James is sounding the alarm. In a recent interview with WAMC’s Grant Ashley, James discussed how voters can spot AI-generated misinformation and protect themselves from manipulation. James, a Democrat running for reelection, faces Republican challenger Saritha Komatireddy, a former federal prosecutor. Early voting in New York begins October 24, and Election Day is November 3. That means the time to learn how to spot fakes is now. James stressed that the old warning signs, such as strange lighting or unnatural movements, are no longer reliable. The fakes have become “really, really good,” she said. What was once easy to dismiss as a clumsy video is now dangerously convincing, and voters cannot simply trust their eyes and ears. They need to verify information through official sources, slow down before reacting, and understand that the technology has advanced faster than most people’s ability to detect it. The stakes are enormous. If a single convincing deepfake can convince even a small group of people to stay home, it can change the outcome of a close race. That is why James is urging every New Yorker to be skeptical, to question what they see, and to rely on official channels for voting information. The core message is simple: in a world where seeing is no longer believing, the best defense is an informed and cautious voter.
James first released a voter protection guide in 2024, but this year’s updated version comes at a critical moment. During the interview, she explained that artificial intelligence can now create fake videos, photos, audio recordings, and even messages that appear to come from your local board of elections. The updated guide is designed to help New Yorkers spot, avoid, and report election misinformation generated by AI. What makes the new guide so important is its warning that the traditional signs of a deepfake are disappearing. It is no longer enough to look for blurry edges or robotic voices, because AI has gotten much better at mimicking real people in real situations. James described how AIs can now produce fake messages that look like they came from official government sources, making it even harder for ordinary voters to know what to trust. One of the most concerning developments is the rise of AI chatbots. More and more people are asking chatbots where and how to vote, and chatbots can confidently provide answers that are completely wrong. James said her office has seen chatbots give voters incorrect deadlines, wrong rules for city or state, and voting requirements that do not exist. This is particularly dangerous because people who are trying to do the right thing can be misled by a machine that sounds authoritative. The best way to avoid this trap, James said, is to call your local board of elections, contact a local elected official, or reach out to the Office of the Attorney General. She also shared the election protection hotline number: (866) 390-2992. Her message was clear: don’t let a chatbot or a deepfake be your final source of truth. Use them as starting points, but always verify with someone who actually knows the rules. When it comes to voting, there is no substitute for official information, and that simple habit could be the difference between casting a ballot and being tricked into staying home.
So what can voters actually do? James offered practical advice that goes beyond the usual warning to be careful. First, look at the source of the information. Does the message come from an established news outlet, a local board of elections website, or a trusted elected official? If not, treat it with suspicion. Second, be aware of the emotional pull. AI-generated misinformation is often designed to trigger strong feelings—shock, anger, fear, or excitement—because emotional content is more likely to be shared and believed. If a message makes you feel outraged or panicked, or if it pushes you to act immediately, that is a red flag. James explained that fakes are designed to get a reaction. They might tell you to act fast, stay home, slow down, or change your plans. They are not meant to inform you; they are meant to manipulate you. The best response is to pause and independently verify anything that seems alarming or urgent. That means closing the app, opening a browser, and going directly to a government website. It means calling your local board of elections if you have any doubt. It means reaching out to the attorney general’s office or a local elected official to confirm what is real. James stressed that there is no substitute for checking with an official source. The updated guide gives New Yorkers practical steps to do exactly that. It teaches people how to question what they see, how to double-check what they hear, and how to report anything suspicious. In a world where fake content can look exactly like the real thing, verification is not just a good idea; it is an essential part of protecting democracy. Voters who take a few extra minutes to confirm information before acting are not being paranoid. They are being responsible citizens. And they are making it harder for bad actors to succeed.
Ashley asked James about the people behind these deceptive AI campaigns and what they hope to gain. James was direct: vote suppression. Bad actors use AI to spread lies more easily and on a much larger scale than ever before. She pointed to a significant example from 2024, when an AI-generated recording that sounded like President Biden called New Hampshire Democrats and told them not to vote. Her office was also involved in a case where a bad actor surgically targeted Black people in Upstate New York, Michigan, and Illinois. That individual gave communities of color false information about where and when to vote. The case was successfully prosecuted in Michigan, with help from Michigan Attorney General Dana Nessel. James explained that this kind of targeted misinformation is especially troubling because it exploits trust and deepens existing barriers to voting. Communities that have historically faced obstacles at the ballot box are being singled out by people who want to keep them from participating. The lies are not random. They are crafted to create confusion, fear, and doubt. A person might receive a text message saying their polling place has changed, a robocall saying their vote does not matter, or a video showing a candidate saying something they never said. Each one is designed to make voting harder or less appealing. James noted that there are many different factors involved in misinformation and disinformation, but the pattern is consistent: suppress the vote by making people feel uncertain, angry, or disconnected from the political process. The scale of this threat is unprecedented. With AI, a single bad actor can generate thousands of convincing messages in minutes. They do not need to knock on doors or pay for expensive advertising. They just need technology, a target list, and a willingness to lie. That is why voter education and reporting are so important. The more aware people are, the less effective these tricks become.
New York has also taken legal action to fight back. James explained that state law now makes it illegal to send political messaging that is AI-generated or manipulated unless it is clearly labeled. The law requires a visible disclosure when content has been created with artificial intelligence. District attorneys’ offices are responsible for prosecuting violations, and the consequences can range from fines to probation to jail time. But enforcement is not always straightforward. Bad actors often hide behind anonymity or operate from outside New York, which makes it difficult to find them and hold them accountable. James said her office is already seeing more fake official notices and manipulated election videos, images, and audio without labels. Many bad actors are simply ignoring the law, and that is why public awareness matters so much. The legal system cannot catch every fake, but voters can help by reporting suspicious content. James urged New Yorkers to contact the attorney general’s office or their local board of elections if they see something questionable. The updated guide includes instructions on how and where to report. Holding people accountable is important, but the larger goal is prevention. By combining legal consequences with an informed public, New York hopes to make AI-generated voter manipulation more costly and less effective. The law sends an important message: deceiving voters is not a clever campaign trick; it is a crime. And while enforcement may be difficult, the state is committed to using every available tool to protect the integrity of elections. The law also gives prosecutors more power to pursue cases involving false voting information, and it gives voters a clear way to recognize when something has been manipulated. Labels matter, and the requirement to disclose AI-generated content is a step toward transparency in a digital world where authenticity is increasingly hard to determine.
Ashley also asked James whether fake AI videos had appeared in her own reelection campaign. James said not as of yet. The answer was calm, but it underscored how quickly this threat can emerge. One day a campaign might be clear of deepfakes; the next day a convincing fake video could be circulating across every platform. As the interview ended, James turned her attention back to the most important remedy: voting. She reminded listeners to go out and vote, and she thanked the host for the opportunity to share the information. The conversation was a powerful reminder that technology will keep evolving, but the fundamental defense against misinformation remains human judgment. No algorithm can force a person to act. No deepfake can cancel a ballot. But AI can try to manipulate, distract, and discourage. The best way to fight back is to be prepared. Voters should talk to friends and family about what they see online. They should help older relatives who may be less familiar with AI and more vulnerable to fake images and messages. They should report suspicious content when they find it. James’s guidance is nonpartisan, even though she is on the ballot herself. Her office’s resources are available to every New Yorker, regardless of political party. The ultimate goal is simple: ensure that every eligible voter can cast a ballot freely, safely, and without being deceived by machines that can imitate anyone. The midterm elections will be a test not only of candidates, but of voters’ ability to see through the noise. With the right tools, a little skepticism, and a clear commitment to democracy, New Yorkers can meet that challenge. The technology may be new, but the answer is as old as democracy itself: stay informed, stay involved, and vote.