The Face of Deception: How AI Images Rocked a Melbourne Council Race and Laid Bare Democracy’s Newest Frontier
It began, as most political scandals do, with a photograph. In the sweltering lead-up to a local government election in Melbourne’s south-east, residents of the Kingston City Council area—covering the bayside suburbs of Mordialloc and Parkdale—began scrolling through their Facebook feeds and pausing. There, on the campaign page of Louise Black, a candidate for the Melaleuca ward, was a series of portraits that seemed almost too perfect. The skin was flawlessly luminous, the lighting sculpted with cinematic precision, and the smile—wide, symmetrical, and impossibly warm—seemed to radiate a synthetic charm. Compared to the candid, slightly disheveled snapshots from her past, these images looked less like photos and more like the renderings of a video game character. For most voters, it was a fleeting curiosity. But for the digital investigators at Deepcut, a media outlet specializing in data and disinformation, it was the smoking gun of a new and troubling era in Australian politics.
When Deepcut ran the flagged images through sophisticated AI-detection software, the results were staggering. The algorithms, which analyze pixel patterns, lighting inconsistencies, and the subtle distortions in facial geometry that plague generative models, returned a probability score of between 96 and 99 percent that the images were manufactured by artificial intelligence. To put that in stark contrast, a genuine photograph of Black posted to her Instagram account in 2024 returned a mere 1 percent likelihood of any AI alteration. There it was, in black and white (or rather, in algorithmic percentages): a candidate for public office had presented a digitally fabricated visage to the electorate. She had effectively run for council with a face that did not exist—or at least, a face that had been meticulously “beautified” by a machine. The question that hung in the humid air wasn’t just if she had done it, but why, and what it meant for the integrity of the ballot box itself.
Dr. Rodrigo Praino, a professor of politics and public policy at Flinders University, offers a chillingly simple explanation for the tactic: it works. “Physical attractiveness has been shown by extensive research to have a clear positive impact on electoral success,” Praino explained to Deepcut. “This is particularly true in the case of low-information electoral contests—that is, when voters typically do not have very large amounts of information on the candidates, their policies, and their records.” Local council elections are the absolute epitome of these low-information contests. Unlike federal or state campaigns, which are saturated with media coverage, policy debates, and party advertising, a council race often hinges on fleeting glimpses of a name on a corflute and a face on a flyer. In that 30-second window of decision-making, a candidate’s perceived attractiveness—their perceived competence, trustworthiness, and health—can be the deciding factor. By using AI to erase blemishes, sharpen cheekbones, and brighten eyes, Black was not just submitting a photo; she was weaponizing the electorate’s subconscious biases.
The election itself unfolded with a high degree of engagement. With a remarkable turnout of just over 79 percent—a figure that far exceeds the national average for local government polls—the residents of Melaleuca ward took their civic duty seriously. When the votes were tallied, however, the AI-enhanced imagery had not propelled Black to victory. She finished in third place on the primary count, securing a respectable but ultimately losing 14.75 percent of the vote. Yet, the drama was far from over. On August 14, following the official declaration of the results, Black and a fellow candidate, Glenn Dumbrell, launched a furious protest. In a joint statement, they demanded a full recount of all ballot papers, citing “serious procedural irregularities” in the counting process. The accusations escalated rapidly, prompting the Victorian Electoral Commission (VEC) to step in. Suspecting that vote tampering may have occurred, the VEC formally referred the matter to the Victorian Civil and Administrative Tribunal (VCAT), throwing the entire outcome of the ward into a state of legal uncertainty. What had started as a story about deceptive marketing had morphed into a full-blown electoral crisis, complete with legal hearings and accusations of fraud lingering over the council chambers.
But the story takes an even darker and more convoluted turn when one considers Black’s political history. This was not her first time contesting the Melaleuca ward. After losing the previous election in November 2024—a mere year earlier—Black found herself entangled in a vicious controversy involving far-right extremism. Anti-fascist researchers and online sleuths alleged that, under various anonymous usernames, Black had posted over 25,000 messages across known neo-Nazi Telegram channels. The accusations painted a picture of a relentless online agitator, hiding behind carefully crafted pseudonyms to spread divisive and racist rhetoric. In one particularly damning exchange, Black allegedly interacted directly with a member of the National Socialist Network (NSN), a prominent neo-Nazi organization in Australia. In the chat, she purportedly asked the NSN member for “poster ideas” for her local campaign, adding a cautionary addendum: “try not to be insanely antisemitic.” The sheer audacity of the request—seeking campaign material from a hate group while simultaneously pleading for a degree of subtlety—laid bare a cynical, transactional approach to politics. It also echoed a similar scandal involving Robbie Beaton, a Liberal candidate for the federal seat of Isaacs in 2022, who was found guilty of providing a false address to the AEC (claiming he lived in the electorate when he actually resided in Camberwell), was slapped with a $1,000 bond, and was also linked to similar antisemitic chat remarks like “bare minimum anti-Semitism.” The parallel between Beaton and Black suggests a disturbing pattern within a certain segment of political aspirants—a willingness to deceive on every front, from their physical appearance to their ideological affiliations.
This tangled web of fake faces and extremist chat logs exposes the yawning chasm between the rapid advancement of technology and the glacial pace of legislative reform. When Deepcut sought comment from the Victorian Electoral Commission regarding the use of AI in candidate imagery, the response was a stark admission of legislative paralysis. A VEC spokesperson confirmed that neither the Local Government Act 2020 nor any associated regulations currently specify how a candidate must present themselves in their identifying imagery. Furthermore, the legislation does not prohibit the use of generative AI to create images. “For this reason,” the spokesperson stated plainly, “how a candidate wishes to use imagery to portray themselves is not regulated by the VEC.” It is a legal free-for-all. In the absence of explicit rules, a candidate can render themselves as a twenty-five-year-old supermodel, place that image on the ballot materials, and face no legal consequence. The law, designed for a world of film cameras and physical photographs, has simply not caught up to the world of deepfakes and diffusion models.
The public, however, is acutely aware of the threat. Professor Nicholas Biddle from the Australian National University (ANU) has been tracking public sentiment on this exact issue, and his findings reveal a population on edge. According to Biddle’s research, people who are more familiar with AI technology are significantly more likely to be concerned about its potential to spread misinformation. The data from the ANU study paints a grim picture of the current information ecosystem: nearly half of all Australians report having seen false or misleading information about candidates and political parties online, and more than four in ten Australians believe that AI will make elections fundamentally less fair in the future. Most tellingly, Biddle notes that a substantial portion of the population already assumes that “there is a substantial amount of use of AI” in political campaigning—a cynical but realistic assessment of the modern digital landscape. Biddle argues that the most critical step toward restoring trust is transparency. “The most important thing is disclosure,” Biddle told Deepcut, “and making clear when AI is being used and when it’s not being used and under what scenario.” Without mandatory labeling, he suggests, voters are left to guess which images are real and which are digital fiction, eroding the very foundation of informed consent that democracy relies upon.
So, what does the future hold? Will voters penalize candidates who are caught with their digital hand in the cookie jar, or will AI-enhanced campaigning become the standard operating procedure? Professor Praino offers a nuanced perspective on the potential backlash. “In the short term, I think we are seeing a lack of trust in the public,” he says, reflecting on the immediate outcry against Black. “But long term, it is harder to tell.” He points to a silver lining hidden within the public’s growing savvy. “There is a lot of awareness about AI in the community. Voters are very aware now that just because a photo is out there, it doesn’t mean it is true. In that sense, an increased awareness could see a positive effect over the longer term.” The fear is that this awareness breeds apathy; the hope is that it breeds skepticism. As the VCAT hearings loom over the Kingston council race, and as the investigation into Black’s alleged hate-speech chats continues, the case serves as a jarring microcosm of a global dilemma. The electorate is now armed with the knowledge that what they see is not always what they get, that the face smiling from the campaign brochure might be a fiction crafted in a server farm, and that the words spoken in secret chats might contradict the sanitized persona on the public stage. We will have to wait and see, as Praino puts it, whether the public’s heightened awareness will translate into a robust defense of authenticity. But one thing is painfully certain: the era of innocent campaign photography is over. Louise Black, whether she wins her recount or not, has permanently altered the landscape of Australian local politics, forcing every candidate and every voter to look at the picture on the billboard and ask a simple, haunting question: Who, or what, am I actually voting for?