A school board member uses AI to portray two colleagues as witches: a case that previews local politics in the generative era

By WWAY TV3 · October 29, 2026.

David Perry, a member of the New Hanover County school board (North Carolina, US), created an AI-generated Halloween image and posted it on a website. According to WWAY, it shows his fellow board members Pat Bradford and Josie Barnhart as witches stirring a cauldron labeled "New Hanover County Schools Poison Potion," with the slogan "Vote Out the Witches." Both are running for reelection.

The board took up the matter again on Tuesday at the end of its meeting, during a broader discussion of board members' conduct. Perry left the room, saying he had business in Concord. According to the station, he later called it "a fun Halloween thing," though he made clear he does not think his colleagues are good representatives. Barnhart replied that she does not see it "as a joke at all."

The episode is not an isolated one. Earlier this month, according to WWAY, the board voted to censure Perry and remove him from his committees until the new board is sworn in in December. The station cites earlier controversies, including alleged breaches of confidentiality and inappropriate language. Perry says he has two years left in his term and will not change the way he operates.

Our take: the underlying conflict is political and personal, and AI is merely the tool. But the tool matters. A few years ago, a poster like this would have required a designer or some skill; today anyone can produce one in minutes at almost no cost. That democratizes expression, including political satire, which has a long, legitimate and protected tradition.

The flip side is that the same ease lowers the barrier for ad hominem propaganda and can easily escalate toward the worst cases: realistic fake images, impersonation or smear campaigns. Here the poster is an obvious caricature, not a deception, and it's worth not overstating things: this is not disinformation but bad taste or provocation, depending on who's judging. That line between obvious satire and plausible falsehood is exactly what future rules will have to draw carefully.

More broadly, local politics is the testing ground for these practices: there is less scrutiny, fewer communications advisers and fewer rules on labeling AI-generated content than in national campaigns. In the short term, expect more cases like this. The sensible response is not to ban the tool but to strengthen codes of conduct, transparency about where images come from, and media literacy.

In the long run, Zendoric's thesis remains optimistic: a technology that lets anyone create at professional quality is a lever for creative abundance. That some people use it first for a school board feud is, no doubt, the noisy price of that democratization, not an argument against it.