Scientists Need a Positive Vision for AI (spectrum.ieee.org)

Security scholars Bruce Schneier and Nathan E. Sanders argue that scientists must reclaim a positive, actionable vision for AI rather than cede the future to pessimism. They note real harms—deepfakes, misinformation, labor exploitation in data labeling, massive energy consumption, military applications, and industry consolidation—and cite mixed expert sentiment (a Pew study showing 56% of AI authors predict net positive effects versus broader scientific worry in an Arizona State survey). The danger, they warn, is that researchers who view AI as a lost cause will disengage, forfeiting the chance to shape its trajectory. To counter this, the authors outline a constructive agenda for the research community: celebrate and scale beneficial applications (e.g., large language models reducing language barriers—including under-resourced sign and indigenous languages—AI-assisted civic deliberation, LLMs for climate communication, national labs building foundation models, and machine learning breakthroughs like protein-structure prediction recognized by the 2024 Nobel). They call for four actions: reform industry practices toward ethics, equity, and trust; document and resist harmful uses; responsibly deploy AI to serve communities; and renovate institutions (universities, societies, democratic bodies) for AI’s impact. Scientists, being close to the technology, have the authority and responsibility to steer AI toward public benefit—outcomes will hinge on choices made today.