Meta has given its artificial intelligence model direct access to users' messages in its Messenger app so that it can detect fraudulent messages and delete them automatically without them reaching the user, according to a report by the American tech website Digital Trends.
This feature is part of a range of new features launched by the company to combat and stop fraud before it reaches users, including an alert that appears to users when they receive a friend request from an account that appears to be fraudulent.
Artificial intelligence can also be made to review any incoming message and suspect that its content is fraudulent or intended to defraud users.
The company is also expanding its efforts to verify its advertisers’ accounts, hoping to make 90% of the ads displayed on the platform verified from verified accounts, making it extremely difficult to post fraudulent ads on the company’s platforms.
Meta stopped 159 million fraudulent ads last year on Facebook and Instagram (Getty Images)
The new features include detailed alerts when attempting to link WhatsApp accounts directly to new devices, in the hope of countering fraudulent attacks in which a hacker links the victim's account to their own devices, according to a separate report from the American tech website TechCrunch.
The report indicates that Meta removed more than 159 million fraudulent ads in the past year alone, 92% of which were removed before the ads even started to run, in addition to removing more than 10 million accounts linked to fraudulent activities on Instagram and Facebook .
Although Meta's new steps aim to reduce the rate of fraud across its platforms, privacy experts are concerned about its AI reading users' private messages.
It is worth noting that Meta previously faced a lawsuit for accessing users' encrypted messages via the WhatsApp platform, according to a previous report by The Guardian.
