Image by Alan Warburton / © BBC / Better Images of AI / Social Media / CC-BY 4.0
J. Nathan Matias, “Humans and algorithms work together — so study them together,” Nature 617, 248-251, 2023. doi: 10.1038/d41586-023-01521-z
“Humans and algorithms work together — so study them together” by J. Nathan Matias explores the need for effective governance of adaptive algorithms that evolve based on user behavior. The author highlights a significant legal case involving Google-owned YouTube’s recommender algorithms, which were implicated in the radicalization of terrorists who killed Nohemi Gonzalez, a student who was tragically killed by ISIS terrorists in Paris. The imminent court ruling on this case has profound implications, as it could hold digital platforms accountable for their algorithms’ recommendations. The paper stresses the urgency of understanding and regulating the impact of these algorithms and proposes interdisciplinary research and collaboration to address the challenge.
The paper further discusses the broader implications of adaptive algorithms, citing instances where their impact has extended beyond radicalization. It mentions how these algorithms played a role in the 2010 stock market crash, exacerbating a mass stock sell-off initiated by humans. Additionally, adaptive algorithms on professional social networks and generative AI programs have affected job transitions and learned to mimic racist hate speech, respectively. To effectively regulate these algorithms and ensure their safe and beneficial integration into society, the author calls for the development of a science that studies collective patterns of human-algorithm behavior. This science would involve breaking down disciplinary boundaries, classifying, explaining, and forecasting these patterns, and working collaboratively with communities affected by algorithms to develop interventions that prevent problems.
Read the full paper here: 10.1038/d41586-023-01521-z