Meta Tests Facial Recognition to Fight Scam Ads and Make Account Recovery Easier

Meta is expanding its use of facial recognition technology to combat scam ads and make account recovery easier. The company announced this on Monday, stating that these tests aim to bolster its existing anti-scam measures and make it harder for fraudsters to exploit Facebook and Instagram users.

Meta Tests Facial Recognition for Combatting Scam Ads

1. Celeb-Bait Ads

   Meta is using facial recognition to detect "celeb-bait" ads, where scammers use images of celebrities or public figures to lure users into clicking on bogus ads. The system compares the faces in ads with the public figure's Facebook and Instagram profile pictures, and if a match is found, the ad is blocked.

2. Facial Recognition for Ad Verification

   Meta claims that the facial recognition technology is only used to fight scam ads and is not used for any other purpose. The company deletes any facial data generated during the ad verification process immediately, regardless of whether a match is found.

3. Protection for Public Figures

   Early tests involving a small group of celebrities and public figures have shown promising results in detecting and blocking scam ads. Meta will soon start displaying in-app notifications to public figures who've been targeted by celeb-bait ads, informing them that they're enrolled in the system. Public figures can opt-out of the system in their Account Center.

4. Detection of Deepfake Scam Ads

   Meta believes that facial recognition can effectively detect deepfake scam ads, where generative AI has been used to create fake images of famous people. This technology can help prevent scammers from using deepfakes to deceive users.

5. In-App Notifications for Public Figures

   Meta will be notifying public figures who are enrolled in the system, giving them the option to opt-out.

6. Opt-Out Option for Public Figures

   Public figures can opt-out of the facial recognition system at any time in their Account Center.

7. Facial Recognition for Identifying Impersonators

   Meta is also testing the use of facial recognition to identify impersonator accounts. This involves comparing profile pictures on a suspicious account with a public figure's Facebook and Instagram profile pictures.

   

Account Unlocking with Video Selfies

1. Video Selfie Verification for Account Access

   Meta is testing the use of facial recognition in video selfies for account unlocking. This method involves uploading a video selfie that is compared against profile pictures on the account to verify identity.

2. Comparison to Phone Unlocking Methods

   Meta claims that this method is similar to unlocking a phone or accessing apps using facial recognition, such as Apple's Face ID.

3. Data Security Measures

   Meta assures users that video selfies are encrypted and stored securely, and are not visible to anyone else. Facial data is deleted immediately after the comparison, regardless of whether a match is found.

4. Potential for Digital Identity Expansion

   By encouraging users to upload video selfies, Meta could potentially expand its offerings in the digital identity space.

No Tests in U.K. or EU

1. Global Tests, Excluding EU and U.K.

   Meta is conducting its facial recognition tests globally, but not in the U.K. or the EU, where comprehensive data protection regulations are in place.

2. Data Protection Regulations in EU

   The EU's data protection framework requires explicit consent from individuals for the use of biometrics for ID verification.

3. Potential for Data-Processing-for-AI in Europe

   Meta's tests in other countries could be a way to pressure European lawmakers to relax data protection regulations for AI purposes.

4. Engaging with U.K. Regulators

   Meta claims to be engaging with U.K. regulators, policymakers, and experts as it moves forward with its tests.

5. Distinction between Security and Commercial AI

   The use of facial recognition for security purposes, like account unlocking, might be acceptable under existing data protection rules, but using data to train commercial AI models raises different concerns.

Summary:

• Meta is testing facial recognition to combat scam ads, particularly those targeting celebrities and public figures.
• The technology is also being used to help users unlock accounts that have been compromised.
• Tests are not currently taking place in the UK or EU due to data protection regulations.
• Meta claims to be engaging with regulators and experts to address concerns about data privacy and the use of facial recognition.