The billion-dollar rivalry over 'ChatGPT for doctors' just got nastier with dueling lawsuits (www.businessinsider.com)

Doximity has filed a federal countersuit against Sequoia‑backed OpenEvidence, escalating a bitter legal fight between two heavyweights racing to build a “ChatGPT for doctors.” OpenEvidence — a $3.5B startup that raised $210M in July — sued Doximity in June, alleging the public company ($13B market cap) impersonated physicians and used prompt‑injection attacks to extract trade‑secret prompts and code. Doximity responded this week by accusing OpenEvidence of false advertising, employee poaching and reputational sabotage, and moved to dismiss the original complaint as speculative. The volley includes claims about OpenEvidence's boast of a “perfect 100%” USMLE score and assertions it never hallucinates, which Doximity disputes with user reports of errors; both sides also trade accusations about mishandling physician identifiers and patient data. The disputes matter beyond the companies: courts could set new precedents on whether using adversarial prompts or account impersonation to probe models constitutes trade‑secret theft. The lawsuits spotlight core technical risks in healthcare AI — prompt injection (malicious inputs that override model behavior), model hallucinations, and potential exposure of system prompts or patient data — and how legal frameworks will adapt. Outcomes will influence product design, defense strategies (rate‑limiting, guardrails, prompt obfuscation), M&A moves (Doximity bought Pathway for $63M), and commercial claims about accuracy in a market where regulatory and reputational stakes are high.