Core Viewpoint - The article discusses the evolution of black and gray market operations due to AI advancements, highlighting the challenges in internet security and risk management as AI becomes a significant player in these illicit activities [1]. Group 1: Black and Gray Market Operations - AI is reshaping the way black and gray markets operate, with predictions that over 50% of information encountered daily will come from AI in the next 5-10 years [1]. - The cost of black market equipment has significantly decreased, with devices like mechanical arms costing around 1,000 yuan and smartphones priced between 300 to 400 yuan [1]. - The underground economy primarily profits from two types of activities: fraud and rebate cheating, with fraudsters needing extensive social outreach to identify targets [1]. Group 2: Resources Supporting Black and Gray Markets - The foundation of these operations relies on three key resources: accounts, devices, and networks [2]. - Accounts are primarily obtained through "cat pools" and "card pools," with a single "cat pool" capable of connecting approximately 512 cards for various illicit activities [2]. - The emergence of cloud phones has transformed the landscape, allowing operators to rent devices at low costs instead of purchasing physical hardware [2]. Group 3: AI's Role in Fraud and Risk Management - Generative AI is being actively utilized by black and gray market operators to enhance their strategies, such as automating customer service interactions to exploit promotional offers [3]. - The concept of "prompt attack" has emerged, where simple keyword prompts can be used to extract benefits from e-commerce platforms [3]. - Companies are countering these threats by employing AI for deep behavioral analysis and creating "AI blue teams" to simulate attacks and improve defenses [3][4]. Group 4: Evolving Risk Management Processes - Traditional content review processes involving both machine and human audits are becoming inadequate due to rising costs and the volume of harmful content [3]. - The risk management workflow is transitioning to a model that incorporates AI machine audits, large model reviews, and expert decision-making to enhance risk assessment capabilities [4].