OpenAI launched Advanced Account Security (AAS) on April 30, 2026, a set of opt-in protections for ChatGPT users designed to guard against unauthorized account access. The program is available to all users but is specifically aimed at high-value individuals such as journalists, political dissidents, researchers, and elected officials.
As part of the initiative, hardware security key maker Yubico announced a partnership with OpenAI to offer two co-branded products — the YubiKey C NFC and the YubiKey C Nano — that can be linked directly to ChatGPT accounts. Security keys are small physical devices that connect via USB and store a unique cryptographic identifier, ensuring only the person holding the key can log into a connected account.
The partnership is intended to address a growing threat of phishing attacks targeting chatbot users. “Ultimately, our intent is to drastically reduce the threat of unauthorized access to sensitive data in OpenAI accounts worldwide,” Yubico CEO Jerrod Chong said in a press release announcing the deal.
OpenAI noted that enterprise users may also benefit from the added protection, given that sensitive corporate information is frequently discussed within ChatGPT sessions. Cybercriminals increasingly target chatbot users in search of extortion-worthy material, according to the source.
The AAS announcement follows a separate OpenAI disclosure that it is launching a new framework for digital defense, and comes weeks after competitor Anthropic announced a cybersecurity model called Mythos — suggesting that security is becoming a growing focus across the AI industry.
There is a notable tradeoff to the new system: if a security key is lost, OpenAI will not be able to help recover access to the account, meaning chat histories could be permanently inaccessible.
Source: TechCrunch