Bug bounty platform Immunefi banned 15 users for submitting ChatGPT-generated bug reports, citing that the AI tool cannot identify genuine bugs and wastes time for all parties involved.
Smart contract bug bounty platform Immunefi banned 15 people for allegedly submitting bug reports created by the generative AI tool ChatGPT. The platform, which rewards whitehats for finding problems with smart contracts in DeFi projects like Aave, Compound, and Synthetix, has paid out over $65 million in bounties by September 2022 with an additional $138 million available for future payouts. Immunefi stated that ChatGPT lacks technical capability beyond providing answers to human inquiries and cannot identify genuine bugs. The platform explained that ChatGPT-generated reports often appear well-written but contain nonsense when analyzed further, wasting time for whitehats, Immunefi, and the projects themselves. According to Immunefi developers, there are obvious patterns in ChatGPT output that make detection possible, such as long, well-structured reports that have no relation to a project's smart contracts. The platform emphasized that genuine bug identification is 'really an art' requiring careful crafting by humans. While Immunefi acknowledged that any genuine bugs highlighted by AI tools should be reported through proper channels, they will continue monitoring for AI-generated submissions and may consider unbanning AI use if such tools become capable of generating real bug reports in the future.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Using AI systems to gain a personal advantage over others such as through cheating, fraud, scams, blackmail or targeted manipulation of beliefs or behavior. Examples include AI-facilitated plagiarism for research or education, impersonating a trusted or fake individual for illegitimate financial benefit, or creating humiliating or sexual imagery.
Human
Due to a decision or action made by humans
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed