OpenAI is stepping up account protection with a new Advanced Account Security programme for ChatGPT users. Partnering with Yubico, the company is introducing hardware-based login keys to combat rising phishing risks, signalling a broader push across the AI industry to secure sensitive conversations and user data.
OpenAI is tightening the screws on account security as concerns grow over the risks tied to increasingly personal and high-stakes chatbot interactions.
The company has introduced a new opt-in feature called Advanced Account Security (AAS), aimed at offering stronger protection for ChatGPT users. While the programme is positioned with high-risk individuals in mind, it is available to anyone looking to add an extra layer of defence to their account.
At the centre of this rollout is a partnership with Yubico, a digital security firm known for its hardware authentication devices. Together, the two companies are launching co-branded security keys that can be linked directly to ChatGPT accounts, helping users guard against phishing attacks.
STORY CONTINUES BELOW THIS AD
The devices, called the YubiKey C NFC and YubiKey C Nano, function as physical login tools. Instead of relying solely on passwords or one-time codes, users must possess the key itself to access their account.
Each device contains a unique cryptographic identifier, making it significantly harder for attackers to gain unauthorised entry even if login credentials are compromised.
The move reflects growing anxiety around chatbot security. As AI tools become more deeply embedded in daily workflows, they are increasingly used to store sensitive information, ranging from personal conversations to corporate data. This makes them an attractive target for cybercriminals, particularly those employing phishing tactics to trick users into handing over access.
OpenAI has suggested that AAS could be particularly useful for journalists, researchers, political figures and others working in sensitive environments. However, the implications extend beyond individuals. Businesses that rely on ChatGPT for internal processes may also see value in stronger account protections, especially as concerns around data privacy intensify.
Yubico chief executive Jerrod Chong said the goal of the partnership is to significantly reduce the risk of unauthorised access to user accounts worldwide. The emphasis on hardware-based authentication marks a shift towards more robust, tamper-resistant security methods.
The launch also comes amid a broader industry push to address cybersecurity challenges in AI. Rival firm Anthropic recently unveiled its own security-focused model, underscoring how protection is becoming a competitive priority alongside performance.
STORY CONTINUES BELOW THIS AD
However, the added security does come with a caveat. If a user loses their physical key, account recovery may not be possible, potentially locking them out of their data permanently.
For OpenAI, the message is clear: as AI systems become more powerful and more personal, safeguarding access is no longer optional.
First Published:
May 01, 2026, 12:36 IST
HomeTechOpenAI rolls out advanced security for ChatGPT with hardware key supportEnd of Article