Visible to the public CAPTCHA: Machine or Human Solvers? A Game-Theoretical Analysis

TitleCAPTCHA: Machine or Human Solvers? A Game-Theoretical Analysis
Publication TypeConference Paper
Year of Publication2018
AuthorsLi, Z., Liao, Q.
Conference Name2018 5th IEEE International Conference on Cyber Security and Cloud Computing (CSCloud)/2018 4th IEEE International Conference on Edge Computing and Scalable Cloud (EdgeCom)
Date Publishedjun
Keywordsautomatic programs, bitcoins, blockchain, CAPTCHA, CAPTCHA solvers, CAPTCHA system, captchas, cheap labor, Complexity theory, composability, Computer crime, computer security, cryptocurrency, Economics, equilibrium analysis, game theoretical framework, game theory, game-theoretical analysis, Games, Human Behavior, Human Solver, human solvers, Machine Solver, open Web resources, pubcrawl, security of data, Servers, traditional wisdom, ubiquitous defense, Web sites, welfare-improving CAPTCHA business model
AbstractCAPTCHAs have become an ubiquitous defense used to protect open web resources from being exploited at scale. Traditionally, attackers have developed automatic programs known as CAPTCHA solvers to bypass the mechanism. With the presence of cheap labor in developing countries, hackers now have options to use human solvers. In this research, we develop a game theoretical framework to model the interactions between the defender and the attacker regarding the design and countermeasure of CAPTCHA system. With the result of equilibrium analysis, both parties can determine the optimal allocation of software-based or human-based CAPTCHA solvers. Counterintuitively, instead of the traditional wisdom of making CAPTCHA harder and harder, it may be of best interest of the defender to make CAPTCHA easier. We further suggest a welfare-improving CAPTCHA business model by involving decentralized cryptocurrency computation.
Citation Keyli_captcha:_2018