People watching TV on platforms like HBO Max and Tubi recently noticed something surprising during commercial breaks. Prominent television personalities and Hollywood actors appeared on screen, recommending online casino apps and promising instant cash payouts. The video footage looked crisp, the voices sounded convincing, and the ads aired inside legitimate streaming services.
Yet none of those celebrities ever agreed to promote those apps. The clips were created entirely with generative software, stitching together stolen video clips and AI-cloned audio to make unauthorised endorsements.
This is not just happening on social media feeds. It has moved onto connected TV screens, slipping straight past the ad vetting systems built to protect viewers.
How do AI celebrity videos get onto major streaming services?
Streaming platforms rely heavily on automated systems to review and approve thousands of commercial submissions every day. Advertisers upload their video files, and automated algorithms scan the content for policy violations, dangerous links, or copyright flags.
The problem is that automated ad checkers were designed to catch traditional violations. They search for known spam signatures, explicit content, or copyrighted music files. When a scammer uses generative software to produce a brand-new video clip, the file has no matching signature in any blacklist database.
The software clones a voice from public interviews and maps it onto video frames of the speaker's face. To an automated reviewer, the video appears to be a standard high-definition commercial. Because the ad buyers pay real money to run the campaign, the systems clear the commercial and serve it directly to subscribers.
What did the evidence show?
Lucid checked the claim that AI-generated video ads using celebrity likenesses to promote casino apps are bypassing ad review systems on major streaming services. The check returned a verdict of Credible.
A detailed report by Inc.com documented how viewers on services such as HBO Max and Tubi began spotting these commercials during regular broadcasts. The issue is part of a broader wave of manipulated promotional content. News organisations including France24, the BBC, CBC, and the Washington Post have reported on manipulated video ads featuring figures like Oprah Winfrey, Tom Hanks, and prominent news anchors.
Regulatory bodies have also highlighted the challenge. The UK Advertising Standards Authority (ASA) noted a sharp rise in unauthorised celebrity endorsements used for online scams. In addition, the Meta Oversight Board recently overturned a decision that had originally allowed a manipulated gambling ad to remain online, pointing to systemic gaps in how platforms detect altered video content.
Why are these commercials so convincing?
Generative video technology has advanced rapidly over the past year. Earlier attempts at digital cloning produced rigid faces, unnaturally glossy skin, and robotic speech cadences that were easy to spot. Today, the tools match lip movements to audio scripts with high accuracy.
Scammers deliberately choose public figures with high trust ratings. Viewers associate recognisable news presenters and well-known actors with authority and truthfulness. When a familiar voice tells an audience that a gambling app offers guaranteed returns, viewers assume the platform hosting the ad checked the credentials first.
Furthermore, these campaigns exploit platform trust. Seeing an advertisement on a television screen carries a higher level of perceived legitimacy than encountering a random post on a social media timeline. Viewers naturally assume that TV networks and streaming apps exercise strict editorial control over everything they broadcast.
How can you spot a manipulated celebrity endorsement?
You do not need specialised software to spot the warning signs of an AI-manipulated video. Here are a few practical habits you can use right away:
Watch the mouth and teeth. Lip sync software often struggles to generate natural tooth movement and soft lip transitions. If the mouth appears slightly blurry or out of time with the spoken words, the video has likely been altered.
Listen to the audio rhythm. AI voice clones often lack natural cadence, taking unnatural pauses between sentences or pronouncing local brand names strangely.
Consider the offer. A major Hollywood actor or news anchor will almost never endorse an offshore gambling application or promise guaranteed financial returns. If the claim sounds unrealistically generous, it almost certainly is.
Check official channels. Before trusting a commercial, search for the celebrity's official social media profiles or website. Public figures frequently publish warnings when their likeness is being used without permission.
How can Lucid help?
When you run across a commercial, offer, or news clip online that feels off, you can check it instantly in the Lucid app. Paste a link, upload a screenshot, or submit a question, and Lucid runs a real-time check against reliable sources to give you a clear verdict with clear reasoning.
Sources
- Inc.com: HBO and Tubi Viewers Spot AI-Generated Celebrity Ads
- France24: How AI clones of celebrities promote gambling apps
- BBC News: Piers Morgan and Oprah Winfrey targeted for US scam ads
- CBC News: AI-generated videos look real in online scam ads
- Washington Post: Celebrities warn followers about AI manipulation
- Meta Oversight Board: AI-Manipulated Video Promoting Gambling
- ASA: A year in scams update