Photo by Google DeepMind on Pexels

Artificial intelligence systems are not immune to bias, even when interacting solely with each other. A new study published in PNAS demonstrates that AI algorithms can develop biases against other AIs, raising concerns about the potential for future AI systems to implicitly discriminate against humans. The researchers suggest that these AI-to-AI biases could inadvertently translate into prejudiced outcomes for human users. The study, available at https://www.pnas.org/doi/pdf/10.1073/pnas.2415697122, gained traction online after being shared on Reddit (https://old.reddit.com/r/artificial/comments/1mmfcva/study_shows_ais_display_aitoai_bias_so_future_ai/). This finding underscores the importance of careful design and monitoring to prevent unintended biases from propagating within AI systems.