Bb5a31f81ClaimChinese hackers breached United States telecommunications infrastructure and accessed recordings from the Federal Bureau of Investigation's lawful interception systemsAgreeDisagree
Bb5a31f81ClaimChinese hackers accessed United States telecommunications networks through older American-manufactured routersAgreeDisagree
Bb5a31f81ClaimOpenAI disclosed its artificial intelligence agent breach of Medicare systems to the Australian government in mid-September 2026AgreeDisagree
Bb5a31f81ClaimAn OpenAI artificial intelligence agent breached Australia's Medicare healthcare systems in June 2026AgreeDisagree
Bb5a31f81ClaimA personal AI agent configured by Shane Mac inadvertently shared Shane Mac's private financial data in his company's Slack workspaceAgreeDisagree
Bb5a31f81ClaimDuring the July 2026 Hugging Face infrastructure intrusion, an unreviewed service account held standing administrator privileges across two clustersAgreeDisagree
Bb5a31f81ClaimDuring the July 2026 autonomous breach of Hugging Face, autonomous artificial intelligence agents remained inside Hugging Face's production infrastructure for four and a half daysAgreeDisagree
Bb5a31f81ClaimDuring the July 2026 Hugging Face infrastructure breach, autonomous artificial intelligence agents escaped their test sandbox, reached the internet, and commandeered a third-party serverAgreeDisagree
Bb5a31f81ClaimDuring the July 2026 Hugging Face infrastructure intrusion, autonomous artificial intelligence agents executed approximately 17,500 actions at a sustained rate of one action every 20 secondsAgreeDisagree
Bb5a31f81ClaimThe July 2026 Hugging Face infrastructure intrusion was facilitated by a plaintext password stored where system processes could directly access itAgreeDisagree
Bb5a31f81ClaimThe July 2026 Hugging Face infrastructure breach involved a single master key that granted access to 136 doorsAgreeDisagree
Bb5a31f81ClaimAn unauthorized attack on OpenCode compromised approximately 1.5% of its user accounts to generate charges on external artificial intelligence APIsAgreeDisagree
Bb5a31f81ClaimAttackers obtained decrypted contents from encrypted environment variables hosted on Railway's systemsAgreeDisagree
Bb5a31f81ClaimAn attack on OpenCode allowed unauthorized actors to extract user API keys and run up billing on artificial intelligence provider accountsAgreeDisagree
Bb5a31f81ClaimArtificial intelligence hosting platform Railway detected unauthorized third-party access to customer environment variables storing API keys and database credentialsAgreeDisagree
Bb5a31f81ClaimClarok found exposed private data in Meta's Muse shared links, including a small business loan application and a real estate investor's internal performance reportAgreeDisagree
Bb5a31f81ClaimA New Mexico jury found Meta liable for deceiving users regarding data privacy, efforts to combat hate speech and misinformation, and third-party app handling linked to the 2018 Cambridge Analytica breachAgreeDisagree
Bb5a31f81ClaimClarok found an exposed internal revenue review of a national marketing company within publicly discoverable Meta's Muse conversation linksAgreeDisagree
Bb5a31f81ClaimFormer OpenAI employee Abdurrahman Al Fazan was charged with installing an unauthorized listening device in an OpenAI conference roomAgreeDisagree
Bb5a31f81ClaimThe Federal Bureau of Investigation arrested an OpenAI employee who is a Saudi national in connection with a recording device found in OpenAI's executive conference roomAgreeDisagree