Artificial intelligence is used in decision support systems in health and law. Accountability questions are critical.
AI in Health
- Image analysis (radiology)
- Pathology diagnosis
- Drug discovery
- Health insurance analysis
- Triage
AI in Health Responsibility
- Physician main responsible (human supervision)
- AI producer: software error
- Hospital: inadequate device
- Data provider: bias
AI in Law
- Supreme Court research
- Document review
- Contract analysis
- Arbitration/litigation outcome prediction
- Robot lawyers (limited)
AI Limits in Law
- Judge Cannot decide on AI (non-human decision is prohibited)
- Lawyer cannot trust AI and submit an uncontrolled petition
- "AI made a mistake" is not an excuse
Human Supervision Condition
- EU AI Act: high risk Human supervision is mandatory in AI
- Turkey: KVKK art.11/g (objection to automatic decision)
- Health: doctor last promise
- Law: lawyer/judge last word
Data Privacy
- Patient/client data transfer to AI is controversial
- KVKK + confidentiality obligation
- On-premise AI should be preferred
- Training with anonymous/fake data
Practical Recommendations
- AI always delivers the output to human beings audit
- On-premise AI for sensitive data
- Examine AI's training data (bias)
- Inform the client/patient about the use of AI
Artificial intelligence and professional liability lawyer recommended.