All articles tagged with #ai evaluation
Originally Published 1 day ago — by WIRED
OpenAI is requesting contractors to upload real and fabricated work examples from their jobs to evaluate AI performance against human tasks, raising concerns about data privacy and trade secret risks.
Originally Published 27 days ago — by Stanford HAI
Stanford experts predict 2026 will focus on rigorous evaluation of AI's utility, increased AI sovereignty efforts by countries, advancements in medical and legal AI, a reassessment of AI's impact on society, and a potential 'ChatGPT moment' in biomedical AI, while cautioning against overhyped expectations and emphasizing the importance of human-centered AI development.