AI War Game Study Prompts Calls for Ethical Safeguards
A recent study found that AI models chose to use nuclear weapons in 95% of war simulations, prompting calls for stronger ethical safeguards. Analysts note the findings have broad implications for AI safety and oversight, including in clinical settings where AI is increasingly used for risk stratification and decision support in maternal care.
- The war game study was conducted by Kenneth Payne at King's College London and involved several leading AI models, including OpenAI's GPT-5.2, Anthropic's Claude Sonnet 4, and Google's Gemini 3 Flash. - In maternal health, AI is being developed to predict and provide early warnings for life-threatening complications such as postpartum hemorrhage and preeclampsia by analyzing electronic health records and ultrasound data. - AI-driven tools can also help standardize fetal biometric measurements from ultrasounds, which can reduce variability between operators and help address racial and geographic disparities in maternal healthcare. - The American College of Nurse-Midwives has not yet issued a formal position on AI, but the technology is already being used in midwifery education through AI-driven virtual patients and customized assessments to enhance clinical skills in a controlled environment. - Ethical frameworks for AI in healthcare emphasize the importance of preventing bias, ensuring patient privacy, and maintaining human oversight and accountability for decisions influenced by algorithms. - In Virginia, recent legislation (HB 1904) allows Certified Nurse-Midwives (CNMs) to be part of the 24-hour on-call rosters for nursery care in hospitals when a physician is not available, a change aimed at addressing maternity care shortages. - As of July 2025, Virginia law grants CNMs who have completed 1,000 hours of supervised clinical experience the authority to practice independently without a written practice agreement with a physician. - The U.S. currently has a shortage of about 8,000 obstetricians, a gap that is projected to potentially increase to 22,000 by the year 2050, highlighting the need for an expanded midwifery workforce.
