Challenges and risks with AI

• Bias can affect results - e.g. loan-approval model discriminates by gender
• Errors may cause harm - e.g. autonomous vehicle experiences system failure and causes collision
• Data could be exposed - e.g. medical diagnostic bot trained using sensitive patient data, which is stored insecurely
• Solutions may not work for everyone - e.g. Home automation assistant no audio output for visually impaired users
• Users must trust a complex system - e.g. AI-based financial tool makes investment recommendations
• Who's liable for AI-driven decisions? - e.g. Convicted of a crime based on evidence from facial recognition