Opinion Column: The Growing Threat of AI Deepfakes

By Guest Columnist Michael Watson, Mississippi Secretary of State

In a conference room at a recent cybersecurity training, I found myself squinting at a screen, playing a high-stakes game of “real or fake.” The assignment was straightforward—sort out which images projected in front of us were authentic and which were AI-generated. Most in the room, myself included, struggled to tell the difference.

It was a wake-up call. If a group of AI-informed participants had trouble distinguishing fakes, how easily could these deceptions skew the information creeping into the feeds of Mississippi voters?

As we approach Election Day, it is imperative Mississippians arm themselves with knowledge of the growing threat of AI deepfakes. These hyper-realistic, AI-generated forgeries of videos, images, and pieces of audio aren’t just a sign of an advancing technological landscape—they pose a real risk for our elections.

In today’s digital age, technology has the power to bring us a world of new opportunities but poses great challenges. On one hand, we’ve seen AI open doors to incredible innovation, with the power to transform industries and reimagine how we live and work. This same technology has also birthed sophisticated tools of deception including deepfakes.

According to a recent Microsoft report, state-sponsored actors like Russia and China have already weaponized disinformation to destabilize democracies. In the U.S., deepfakes have targeted candidates from across the political spectrum—Vice President Kamala Harris, former President Donald Trump, and many others down the ballot—turning elections into fertile ground for digital deception. Their goal is to mislead, confuse, and weaken Americans’ faith in our electoral system.

Why should Mississippi voters care? In short, a single deepfake circulated at the right time could mislead thousands of voters, change the outcome of a race, damage trust in elections, and most importantly, causes democracy to falter.

In Mississippi, our team worked with the legislature during the 2024 Legislative Session to pass Senate Bill 2577, which ensures bad actors are penalized for the wrongful dissemination of deepfakes. We have also implemented tools like Elections 101 and the My Election Day online resource as sources of trusted information.

Many tech companies also signed an AI Tech Accord earlier this year to assist the public with detecting and preventing the spread of deepfakes, thus helping voters retain the right to decide who governs them. Innovations like watermarking, content credentials, and advanced detection tools are being developed and deployed to help differentiate the real from the fake.

In the end, the final line of defense is you—the voter. Every voter must take responsibility for the information he or she consumes. Learn to recognize the red flags of deepfakes, like unusual lighting, mismatched audio, and awkward lip movements. Question sensational media and only share reputable content. The onus is on all of us.

As November approaches, remember this: technology may arm us with powerful tools to combat deepfakes, but public awareness remains our strongest weapon. The power to protect our elections lies in our hands. Let’s make sure we use it.