The authors further argue that the deployment of AI technology hurts workers, as it relies on training data taken without compensation and often involves repetitive and traumatic labor for gig workers. They criticize AI-related policy for being driven by corporate interests and junk science, and call for policymakers to focus on solid research investigating the harms and risks of AI. They emphasize the need for regulation to address issues such as data accumulation, climate costs, damage to the welfare state, and the disempowerment of the poor.
Key takeaways:
- The real dangers of AI are not the potential to wipe out humanity, but the existing harms such as wrongful arrests, defamation, and the spread of misinformation.
- AI technology is already enabling discrimination in housing, criminal justice, and health care, and is becoming more prevalent in the workplace.
- AI-related policy must be science-driven and built on relevant research, but much of the current research is nonreproducible and full of hype.
- There is a need to focus on solid scholarship that investigates the actual harms and risks of AI, including the unregulated accumulation of data, climate costs, and the disempowerment of the poor.