Meta has begun testing facial recognition technology to improve security on Facebook and Instagram platforms. The new approach is focused on protecting users from malicious ads featuring celebrity images, and fake accounts and making it easier to regain access to hacked accounts.
Fighting celebrity-bait ads and fake accounts
Ads often use photos of celebrities and public figures to draw users’ attention to the ads. These ads violate Meta’s policies and can lead to sites that collect personal data or require users to transfer money. Such fraudulent ads have led to lawsuits against Meta on more than one occasion:
- In Japan, 30 people sued Meta over fraudulent advertising: https://cpa.rip/en/news/lawsuit-against-meta-in-japan/
- An Australian court ruled against Meta for advertising fraudulent crypto projects using celebrities: https://cpa.rip/en/news/meta-lawsuit-court-order/.
- Japanese billionaire Yusaku Maezawa sues Meta: https://cpa.rip/en/news/japanese-billionaire-sues-meta/
- Meta is being sued over crypto investment scam ads: https://cpa.rip/en/news/fake-investments-ads-on-facebook/
Meta develops and tests facial recognition algorithms that automatically compare images from suspicious ads to celebrity profile pictures on Facebook and Instagram. If there is a match and fraud is confirmed, the ad will be blocked and the data collected for comparison to protect personal information will be deleted immediately after verification.
Meta also plans to use facial recognition technology to detect fake accounts that are masked as celebrity profiles. This can further protect users, especially those who trust public faces, and can easily fall into the trap of scammers.
New system for restoring access to accounts
Meta is also testing the ability to restore access to accounts via video selfies. Users who have lost access to their accounts will now be prompted to upload a short video selfie. It will be used for a one-time comparison with profile pictures to confirm the owner’s identity. Meta assures that the selfie video will be encrypted and deleted immediately after verification, preventing further use.
Transparency and data protection
Meta emphasizes data privacy, assuring that the new protections are being thoroughly tested for security compliance. In the coming weeks, the company will be sending notifications to a wider group of public individuals who suffered from celeb-bait informing them of the bait-ads protection measures. They will be able to opt out of this system via the Accounts Center.
Meta is working with regulators and experts so that these innovations can help protect users while remaining within strict security and privacy standards.
Possible impact on arbitrageurs and advertisers
With the testing of new facial recognition technology, Meta aims to make Facebook and Instagram more secure from fraudulent ads and fake accounts. These measures may already be starting to affect media buyers.
- Meta’s moderation system uses machine learning to analyze images, videos, and text. The introduction of facial recognition will make it even more sensitive to circumvention attempts. Now it will be more difficult to use “gray” promotion schemes with images of celebrities and public faces. This will encourage users to search for more original, transparent, and safe creatives that comply with Meta’s policy.
- The process of restoring profiles in case of blocking will become more complicated, which will probably lead to an increase in the cost of consumables (advertising accounts).
