AI start-up claims Anthropic breach and access to employee data
An AI start-up claimed it hacked Anthropic in 12 hours and gained access to employee data and internal messages. The start-up claimed it compromised an Anthropic AI agent through a supplier tool.
Key facts
- Affected organisation
- Anthropic
- Incident type
- Data breach
- Occurred / detected
- Not confirmed
- Scale of damage
- Claimed access to employee data and internal messages
- Cause / method
- Alleged compromise of an AI agent through a supplier tool
- Location
- United States
If your data was exposed
- Don’t tap links in texts, calls or emails that mention this incident and offer compensation, refunds or breach checks. Check directly on the organisation’s official website or app.
Timeline
- An AI start-up claimed it hacked Anthropic in 12 hours and gained access to employee data and internal messages.
- Claims that an Anthropic AI agent had been compromised through a supplier tool became public.
2 related reports
Other incidents at Anthropic
Anthropic breach exposes data from 112 Claude business accounts in South Korea
Impact: Information from 112 Claude business accounts in South Korea exposed
Unauthorised access to Anthropic's internal systems through a vulnerability in third-party business software exposed information from 112 Claude business accounts in South Korea. The exposed data included names, email addresses, account and organisation identifiers, and usage and spending metrics, while the company said conversations and payment information were not accessed. The Personal Information Protection Commission (PIPC) launched a formal investigation on 10 November, and it was subsequently confirmed that Anthropic reported the intrusion to South Korean authorities 21 days after detecting it and notified customer companies after 28 days.
Related reports: 33 · last report →Claude users hit by login session theft and account hijacking
Impact: Claude user accounts hijacked and usage consumed without authorisation
Anthropic's Claude users suffered information-stealing attacks and login session hijacking. Infostealers stole login sessions, leading to usage the account holders had not made being consumed. On 10 September, users continued to complain of account hijacking.
Related reports: 3 · last report →Customer data leak alleged after Anthropic Claude outage
Claims emerged that customer data was leaked following Claude's service outage on Friday. Anthropic is investigating the claims, and whether a leak actually occurred and its scale have not been confirmed.
Read on Cybernews ↗Anthropic security AI Mythos allegedly hacked
Allegations have been raised that Anthropic's security AI Mythos was hacked. Whether a breach actually occurred, the scale of the damage and the response could not be confirmed.
Read on 지디넷코리아 ↗Employee error at Anthropic leaks 510,000 lines of Claude Code source
Impact: 510,000 lines of Claude Code source code
The source code of Claude Code, Anthropic's AI coding tool, was leaked. The leak amounted to 510,000 lines, and the cause is reported to have been an employee's mistake.
Related reports: 2 · last report →Anthropic's X account hacked to promote fake Claude token
Anthropic's X account was hacked, and scammers used it to promote a fake CLAUDE token. The scale of financial losses and the response, such as account recovery, have not been confirmed.
Read on BeInCrypto ↗Anthropic data leak affects customer accounts
Impact: Customer accounts affected
A data leak at Anthropic affected customer accounts. The date of the incident, the number of affected accounts, the cause of the leak and the response have not been confirmed.
Read on VentureBeat ↗
The summary and key facts are written by AI from related reports and may contain errors. Check the original articles and the organisation’s official statements for accurate details.