US EditionPublished • loading... • Updated
You Can Now Protect Your ChatGPT Account with a Special USB-Key
The opt-in mode uses passkeys or hardware security keys and automatically opts users out of model training, OpenAI said.
- On Thursday, OpenAI launched "Advanced Account Security," an opt-in feature for ChatGPT and Codex that replaces traditional passwords with hardware security keys or passkeys to prevent phishing attacks.
- OpenAI designed this mode to protect high-risk users including journalists and government officials, citing a 2024 report by Group-IB identifying over 100,000 stolen ChatGPT credentials circulating on dark web marketplaces.
- Partnering with Yubico, OpenAI offers a discounted $68 hardware security key bundle; however, users must manage their own recovery keys because the company's support team cannot restore access if credentials are lost.
- Members of the Trusted Access for Cyber program, which grants security researchers access to advanced models, must enable Advanced Account Security by June 1, 2026, or submit alternative phishing-resistant authentication proof.
- This launch signals a broader cybersecurity strategy for ChatGPT's 900 million weekly users, protecting sensitive account data as the platform shifts toward universal passwordless authentication across its ecosystem.
Insights by Ground AI
18 Articles
18 Articles
OpenAI launches hardware security keys for ChatGPT with Yubico partnership and disables password login for high-risk users
OpenAI's Advanced Account Security replaces passwords with hardware keys and passkeys, disables email recovery, and opts users out of model training. Co-branded YubiKeys cost $68 for two.
·Amsterdam, Netherlands (Kingdom of the)
Read Full Article
Las Vegas Sun+4 Reposted by 4 other sources
OpenAI and Yubico Partner to Bring Custom Phishing-Resistant YubiKeys to OpenAI Users
None
·Las Vegas, United States
Read Full ArticleCoverage Details
Total News Sources18
Leaning Left5Leaning Right0Center2Last UpdatedBias Distribution71% Left
Bias Distribution
- 71% of the sources lean Left
71% Left
L 71%
C 29%
Factuality
To view factuality data please Upgrade to Premium
