Iterated crowdsourcing dilemma game.

Area of Science:

• Game Theory
• Computational Social Science
• Crowdsourcing Security

Background:

• Crowdsourcing enables collective problem-solving for complex tasks via the internet.
• The open nature of crowdsourcing poses security risks, including sabotage, theft, and manipulation of solutions.
• Existing models often consider single-round interactions, not the impact of repeated engagements.

Purpose of the Study:

• To analyze the impact of repeated interactions on the crowdsourcing dilemma.
• To identify evolutionarily stable strategies in iterated crowdsourcing games.
• To determine conditions under which repeated encounters improve overall social welfare.

Main Methods:

• Extension of a previously proposed crowdsourcing dilemma game into an iterated game framework.
• Enumeration of pure evolutionarily stable strategies within the class of reactive strategies (dependent on the opponent's last action).
• Analysis of 4096 possible reactive strategies to identify stable ones across different parameter regions.

Main Results:

• Identified 16 pure evolutionarily stable reactive strategies out of 4096 possibilities.
• Repeated encounters can enhance social welfare, but only under specific conditions: low attack damage and low attack cost.
• In most parameter spaces, iterated interactions do not substantially alleviate the crowdsourcing dilemma.

Conclusions:

• The iterated crowdsourcing game framework reveals specific stable strategies for managing security risks.
• The potential for repeated interactions to improve social welfare is highly contingent on the economic parameters of attacks.
• While beneficial in limited scenarios, iterated play does not universally solve the crowdsourcing security dilemma.