University of Washington
3 storiesThe University of Washington is a public research university in Seattle whose researchers take part in AI research and public debate. Recent coverage includes a Nature method co-authored by UW researchers that converts existing LLMs into byte-level models using under 1% of the usual compute; professor Emily Bender calling AI's "existential" risk "fake" and pointing to environmental costs, worker disruption and misinformation; and adjunct professor Chirag Shah saying human extinction is unlikely and nearer-term risks include biased hiring algorithms and unmonitored agentic AI.
Related topics
Researchers from LMU, the Allen Institute for AI, Cambridge, the University of Washington and Imperial College London published a Nature method that converts existing LLMs into byte-level models without retraining from scratch, using under 1% of the usual compute. The byteified models largely keep original performance, add byte-level benefits, outperform prior same-size byte-level LLMs and handle character-level tasks like spelling backwards far better.
University of Washington professor Emily Bender argues the AI safety debate focuses too much on hypothetical existential threats while overlooking present harms such as environmental costs, worker disruption, surveillance, misinformation and chatbot dependence; incidents where AI systems access outside networks should be the responsibility of deploying companies, not autonomous models 'going rogue.'
Former Anthropic researcher Jacob Coxon warned AI could threaten humanity this decade, but UW adjunct professor Chirag Shah calls human extinction unlikely. Shah says nearer-term risks are biased hiring algorithms, unmonitored agentic AI behavior, and erosion of critical thinking and education, and notes the US–China AI race discourages slowing development.
That is everything
The A.I. Daily
Study: Byte-Level LLMs Can Overcome Subword Tokenization Limits
