Signal Paper argues human-centered AI design alone can't stop generative AI from fostering learner dependence

Summary

The paper argues that as generative AI's agentic capabilities expand, learners are promised unprecedented gains in productivity and learning even as they are pushed toward individualism, over-reliance, and dependence. It contends that human-centered design approaches such as value-sensitive design, which assume that surfacing human needs and values is enough to prevent such harms, cannot reach risks rooted in broader structural forces. Those forces include economic incentives and people's inherent tendency to take the easiest path available in the moment. Because of this, the authors argue that refining a system's attention to human values alone will not resolve dependence unless the surrounding economic and behavioral pressures also change. In response, they propose an emancipatory design vision for generative AI aimed explicitly at learner flourishing rather than productivity metrics alone.