A recent study highlights the capacity of artificial intelligence to engage in collective decision-making and exert influence over one another. This capability raises concerns, particularly if AI falls into the wrong hands, according to experts. The study indicates that AI systems can not only make decisions as a group but also sway each other's choices.
AI's ability to simulate and represent diverse groups in decision-making is being explored. AI models can represent future generations and marginalized groups, ensuring their perspectives are included in complex modelling and policy design. However, caution is advised when deploying AI in democratic processes. Factors such as voting methods and presentation order can influence AI voting outcomes, and biases can exist in LLM voting behaviours compared to human voters. Despite these challenges, AI offers innovative ways to enhance collective intelligence by managing vast data and optimising processes.
While AI offers opportunities for inclusive participation and enhanced collective intelligence, responsible design and deployment are crucial. Further research is needed to mitigate biases and ensure that AI integration aligns with democratic values.
