Brain-computer interfaces risk centralising data and extending bias


In September 2018, AI Now co-founder Meredith Whittaker sounded the alarm about the potential for abuse of the convergence of neuroscience, human enhancement, and AI in the form of brain-computer interfaces. Part of Whittaker's concern was that the only companies with the computational power necessary to develop these technologies are those already leading in AI: Google, Facebook, Microsoft, and equivalent. The result would be that the neural data collected from individuals' thoughts would be stored in a centralised server infrastructure controlled by just a handful of very large private companies. Whittaker was also concerned about the potential for bias; in her example, the helmet used by one lab to collect data from the brain works better for men than women, leading to a system that will view men as the norm.

tags: brain-computer interfaces, AI, bias, human enhancement, neuroscience, data mining
Writer: Jeanne Carstensen
Publication: All Turtles blog

See more examples
Related learning resources