X auto-locks first crypto posts, adds selfie checks to curb bots

X activated the “First-Mention Auto-Lock” protocol, which locks any account making its first post with keywords such as “crypto,” “bitcoin,” “airdrop,” or “wallet.” To regain access, users must complete a multi-factor “proof-of-humanity” check that includes a selfie video, and a 12-hour hold applies to replies and reposts.

According to X’s engineering team, the system targets the “Zero-Day Bot” strategy, where automated accounts lie dormant before posting investment scams. By stopping the first crypto-related post from previously silent profiles, the platform aims to curb a pig butchering scam and “drainer” schemes.

In the first 24 hours, figures shared by X indicate more than 1.4 million fraudulent posts were intercepted and nearly 800,000 automated accounts were permanently suspended. The company describes the checks as “biometric barriers” intended to raise the cost of running botnets.

The rollout has drawn criticism from users and civil liberties advocates who argue that locking accounts for mentioning technology chills legitimate speech. In a public thread, the safety lead put the false-positive rate around 1.2% and noted that most real users complete verification in under three minutes. Advocacy groups counter that the protocol enables “content-based account suspension.”

The company positions the protocol as a security-first trade-off designed to restore confidence in the verified checkmark as a sign of human activity. It describes the system as part of a broader 2026 anti-fraud program focused on limiting impersonation and link baiting in high-profile discussions.

X is working with cybersecurity firms to test the integration of “on-chain reputation” signals into the workflow. Under the concept, users with a clean record of blockchain activity could qualify to bypass the auto-lock while checks remain in place for new or inactive accounts. Internally, the company refers to the approach as “Information Finance.”

The platform has not outlined geographic limits and indicates the rollout is platform-wide. X plans to track interception rates, suspension outcomes, time to verify, and attempted evasion, and to refine keyword triggers and biometric review steps based on those results.