An investment ad on Facebook showed Martin Lewis appearing to endorse an opportunity backed by Elon Musk. The video was fake: Lewis had not endorsed the offer, and the footage was an AI-generated deepfake. He warned that the ad was a scam intended to steal money.
The incident brought together two concerns: the use of synthetic media to make a scam more convincing, and the ability of paid social media ads to reach potential victims. Lewis urged people to share the warning and called on government and regulators to act.
A familiar face in a more convincing scam
Scammers had already used Lewis’s image in social media advertising, despite his stated practice of never appearing in adverts endorsing third-party products or services. This time, the ad went further by showing a moving, speaking version of him.
In an interview with ITV’s Good Morning Britain, Lewis described it as the first deepfake scam advert he had seen featuring him. He said the technology was still at an early stage and warned that it would improve. That prospect changes the challenge for viewers: a familiar face and voice may no longer be enough to establish that an endorsement is genuine.
The video was circulating on Facebook. Lewis said he was angry about the ad and had campaigned for years for stronger regulation of scam advertising.
What Meta said it did
Lewis had previously taken legal action against Facebook over scam ads carrying his image. He settled the defamation suit in 2019 after Facebook agreed to changes, including a report ad button for U.K. users and £3 million in funding to help establish a citizens’ scam advice service.
After TechCrunch asked about the deepfake, a Meta spokesperson said the company was investigating. Meta also said the original video had been proactively removed and that it had removed copycat adverts using the same imagery.
That response did not answer TechCrunch’s direct question about how the ad had been allowed onto the platform. The gap between removing an ad and explaining how it passed through a platform’s systems is central to Lewis’s criticism: the fake was already circulating before it was taken down.
Why ad targeting matters
The scam depended on a deceptive message, but the article also points to the role of advertising tools in distributing it. Scammers can place an ad on a platform and use its targeting features to try to reach people they believe may be more susceptible to the pitch.
Those tools are built around behavioral targeting and user profiling. In this case, the concern is that a convincing deepfake combined with targeted distribution could put the offer in front of people likely to respond. Removing a reported ad addresses that instance; it does not, by itself, resolve questions about how such ads are submitted, shown and detected.
The scale of financial losses adds weight to the concern. The article reports that fraud and financial scams cost U.K. consumers and businesses a record £1.3 billion last year. It also notes that Meta’s full-year earnings for 2022 were $116.61 billion, underscoring the difference in scale between the company’s revenue and the contribution Facebook had agreed to make toward scam advice.
Pressure for stronger rules
Lewis criticized the U.K. government’s response, pointing to two consultations on online ads and the time it was taking to pass the Online Safety Bill. The bill was expanded to cover scam ads last year, but was still moving through parliament at the time of the article.
Until the legislation passed, Lewis argued, users remained dependent on platforms’ own enforcement of their terms. His warning was that paid promotion gives technology companies a direct role in distributing these ads. The deepfake therefore raises a broader question alongside the immediate scam: who should be responsible when a platform carries a false endorsement, and what should happen before another viewer sees it?