Next Week in Science, November 13, 2025
James Watson, who died at the age of 97, leaves behind a legacy both towering and deeply complicated. In 1953, alongside Francis Crick, Rosalind Franklin, and Maurice Wilkins, Watson helped uncover the double-helix structure of DNA, one of the most consequential scientific breakthroughs of the 20th century. This discovery transformed biology, clarified the molecular basis of heredity, and paved the way for modern genetics, biotechnology, precision medicine, and forensic science. For decades, his work has shaped humanity’s understanding of life itself.
Watson went on to lead the Human Genome Project in its formative years and served for decades as a prominent figure in the field of molecular biology. His early scientific achievements earned him global recognition, including a share of the Nobel Prize in Physiology or Medicine in 1962.
Yet Watson’s career in later years was overshadowed by repeated, unfounded, and harmful statements on race and intelligence that were widely condemned by scientists, civil rights leaders, and the institutions with which he was affiliated. These remarks, rooted in prejudice rather than evidence, strained his relationship with the scientific community and cast a long shadow over his public legacy.
So, here’s what’s in the news, and the research.
Stylometric comparisons of human versus AI-generated creative writing
A new study using stylometry, specifically Burrows’ Delta, reveals that human-written and large language model (LLM)-generated creative texts like those from GPT-3.5, GPT-4, and Llama 70b are stylistically distinguishable. Despite LLMs’ advances, their outputs show consistent stylistic patterns distinct from the diverse styles of human authors.
California Enacts New AI Safety and Transparency Laws While Vetoing ‘No Robo Bosses Act’
California has introduced new AI regulations that require businesses to provide consumers with notice before using automated decision-making technologies that affect significant decisions, including employment-related ones, starting in 2027. The state has also enacted a law mandating detailed safety disclosures for “frontier” AI model developers with large computing capacities and annual revenues exceeding $500 million, effective as of 2026.
