OpenAI invites Pakistani experts to ‘hack’ ChatGPT for Rs 14 million. Full details here
Pakistani and other Asian cyber security researchers have quietly made a name for themselves in global hacking competitions, bug bounty programs and digital security research. Now, one of the world’s biggest AI giants has opened another door.
OpenAI has officially added Pakistan to the list of eligible countries for its exclusive OpenAI Bio Bounty program, giving talented Pakistani researchers the chance to earn up to US$50,000 to help secure the next generation of ChatGPT. But this is far from a typical bug bounty.
Instead of hunting for software bugs or website vulnerabilities, participants will be asked to challenge one of AI’s biggest security questions: Can the world’s most advanced AI models be tricked into bypassing their biological safeguards? The answer could shape how future AI systems are made secure.
Despite the headline-grabbing prize, it’s no match for casual ChatGPT users. OpenAI is looking for experienced researchers with backgrounds in cyber security, AI read teaming and biological security. To join, applicants must submit a brief application detailing their experience, affiliation and country of residence. An active ChatGPT account is also required.
After screening, OpenAI will invite selected researchers to a private testing environment, where they must sign a non-disclosure agreement before starting their work. Those who previously applied for the program will not need to submit a second application. The real challenge begins when researchers gain access.
Participants will test GPT-5.6 and future AI models in search of what OpenAI calls the Universal Jailbreak, a technique capable of consistently breaking a model’s biological security protections. It’s not about discovering a single prompt that slips through the cracks.
Researchers should instead identify a broader methodology that can repeatedly bypass system considerations across a range of biologically relevant scenarios. If successful, their findings will help bridge these gaps before OpenAI can be used in the real world. To encourage intensive research, OpenAI has increased the program’s maximum award from US$25,000 to US$50,000.
Researchers who uncover a major biosecurity vulnerability — or a universal jailbreak affecting GPT-5.5 or GPT-5.6 — can receive the full amount. The company also says it may make smaller payments for significant partial discoveries. Testing of GPT-5.5 will continue until July 27, 2026, after which researchers will focus exclusively on GPT-5.6.
AI is becoming increasingly powerful, bringing new security concerns as well as immense opportunities. Technology companies are investing heavily to ensure their AI systems cannot be used for malicious purposes, particularly in areas involving biological threats. One of the most effective ways to strengthen these defenses is to invite credible researchers to actively try to break them down. This process, known as AI read teaming, has become an integral part of building secure AI systems.
Pakistan has consistently emerged as a growing force in cyber security, with researchers regularly receiving recognition through international bug bounty programs and security competitions. OpenAI’s decision to include Pakistan in the biobounty program is another sign that local expertise is being looked at globally. For qualified researchers, the opportunity extends beyond a potential $50,000 award.
Chat GPT picks Pakistan as “favourites” ahead of thrilling India T20 World Cup showdown
The post OpenAI invites Pakistani experts to ‘hack’ ChatGPT for Rs 14 million Complete details here appeared first on Daily Pakistan English News.
Stay updated with the latest alerts on this story:
View Original Coverage
