184 million social media logins exposed online; Meta pushes toward fully automated ads
A cybersecurity researcher recently discovered an unsecured database containing more than 184 million unique login credentials for services including Instagram, Facebook and Snapchat, highlighting the growing threat posed by infostealer malware. Security experts recommend three immediate steps: change passwords regularly and never reuse them across accounts, using unique, long and complex passwords stored in a password manager; enable two-factor authentication wherever possible, ideally through an authenticator app rather than text messages; and regularly audit email inboxes for sensitive documents and old passwords.
By 2026 Meta intends to hand advertising over to AI end to end. The company's chief executive says advertisers want AI products producing measurable outcomes at scale, with Meta aiming to offer an AI one-stop shop in which businesses define goals and budgets while the platform manages execution. Across Meta's family of apps there are 3.43 billion unique active users worldwide, and the company's AI ad tools already generate ad variations, image backgrounds and automated video adjustments. Some observers called the announcement the death of the ad agency, though many small businesses and agency owners counter that Meta's current AI ad tools still produce bugs and mistakes.
A Facebook Groups bug is causing headaches for community owners: one admin reported that hundreds of automatic invitations were sent from her new paid group without her knowledge, with people mysteriously joining and invites having to be removed one at a time. Others with similar experiences say the sole fix they found was closing the group and creating a fresh one.
A demonstration video made entirely with Google's new Veo 3 AI tool is circulating as a warning about how fast realistic deepfakes are evolving. Glitches are getting harder to spot, faces look real and movements feel natural, raising the stakes for verifying content before trusting it.
