arXiv threatens one-year ban for AI hallucinations

arXiv, the prominent open-source research repository, announced last week that it would ban scholarly authors from the platform for up to one year if "hallucinated references" are discovered in their work. The policy clarification, made by arXiv computer science chair Thomas Dietterich, states that authors bear full responsibility for all contents of their papers regardless of how they were generated. However, the announcement has sparked significant backlash among researchers, with critics including economics professor James Miller, AI researcher Luca Ambrogioni, and former neuroscientist Neal Amin arguing the policy is overly strict and fails to account for human error in lengthy academic publications. The debate over AI-generated content extends beyond academia into broader discussions about the erosion of quality and craftsmanship. A separate opinion piece argues that while AI is often blamed for declining craft in music, furniture, and other goods, the real responsibility lies with consumers and companies who prioritized speed and cost over quality. The piece contends that automation and digital optimization have already diminished human connection and artistry, with AI simply accelerating an existing trend. The author suggests brands that invest in craft, storytelling, and genuine human interaction can differentiate themselves in a market filled with "blands," though the piece acknowledges AI and measurement tools can serve as useful instruments without becoming substitutes for authentic craftsmanship.