Meta Tests Facial Recognition to Combat Celebrity Scam Ads

Meta has announced a significant expansion of its facial recognition tests aimed at tackling celebrity scam ads. This initiative, revealed on Monday, is part of a broader effort to enhance user safety on its platforms, Facebook and Instagram.

Enhanced Anti-Scam Measures

According to Monika Bickert, Meta’s VP of Content Policy, these tests are designed to strengthen existing anti-scam measures. By integrating advanced machine learning algorithms into its ad review system, Meta aims to prevent fraudsters from exploiting its platforms to mislead users.

“Scammers often use images of public figures, such as celebrities and content creators, to lure people into clicking on fraudulent ads,” Bickert stated.

These deceptive ads, commonly referred to as celeb-bait, not only violate Meta's policies but also pose risks to unsuspecting users who might be misled into sharing personal information or sending money.

Facial Recognition as a Safeguard

The new tests will utilize facial recognition technology to verify the identities of public figures featured in ads flagged by Meta’s systems. Bickert explained, “We will compare faces in ads against the public figure’s profile pictures on Facebook and Instagram.” If a match is confirmed and the ad is deemed a scam, it will be blocked.

Importantly, Meta emphasizes that this feature is solely intended to combat fraudulent ads. Bickert reassured users that “any facial data generated for this one-time comparison will be deleted immediately, regardless of whether a match is found.”

Promising Initial Results

Early tests involving a select group of celebrities have shown promising results in detecting and addressing these scams more efficiently. Meta is optimistic about the potential for this technology to also identify deepfake scams, where generative AI creates misleading imagery of famous individuals.

Despite previous criticisms regarding its failure to curb scams using celebrity likenesses, Meta’s timing for implementing facial recognition measures is notable, especially as the company seeks to enhance user data for training commercial AI models.

In-App Notifications for Public Figures

In the coming weeks, Meta plans to roll out in-app notifications to a broader range of public figures affected by celeb-bait ads, informing them about their enrollment in this protective system. Bickert added, “Public figures can opt-out anytime via their Accounts Center.”

Identity Verification Enhancements

Additionally, Meta is testing the use of facial recognition through video selfies for account recovery. This feature aims to expedite the process for users locked out of their accounts due to scams.

“Video selfie verification will be a quick and easy method for users to regain access to their accounts,” Bickert stated.

The new verification process allows users to upload a video selfie, which will be securely processed and compared to their profile pictures, enhancing security against account takeovers.

Global Testing with Privacy Considerations

While these tests are being conducted globally, Meta has notably excluded the U.K. and EU from this rollout due to strict data protection regulations. This exclusion raises questions about Meta’s approach to privacy as it pushes for broader data processing capabilities in the name of combating fraud.

Andrew Devoy, a spokesperson for Meta, confirmed ongoing engagement with U.K. regulators as testing progresses. “We’ll continue to seek feedback from experts and adapt the features as necessary,” he said.

As Meta navigates these challenges, the distinction between using facial recognition for security versus training AI models remains a critical point of discussion among users and policymakers alike.