California AI law targets opioid crisis as US overdose rates surge

Researchers in California and nationwide are integrating artificial intelligence into the fight against the opioid epidemic. Machine‑learning models analyze smartphone and wearable data to predict relapse risk, while platforms such as NarxCare generate risk scores from prescription histories. In response, California’s Assembly Bill 489, effective Jan. 1 2026, bans AI from posing as a licensed medical professional and requires human oversight of algorithmic decisions.

At the national level, the National Drug Early Warning System reported 1,039 EMS encounters for non‑fatal xylazine‑related overdoses between Jan 2024 and Sept 2025, with the highest rates concentrated in Northeastern counties, especially Pennsylvania. The briefing also highlighted the emergence of a novel research chemical, PipTapentadol, marketed from China. Early user reports on Reddit describe atypical effects and a lack of safety data, raising public‑health concerns.

Separately, Lawrence Livermore National Laboratory developed a random‑forest model that can identify previously unknown synthetic fentanyl analogues by their chemical properties, offering law‑enforcement a reference‑free tool to keep pace with illicit drug modifications.