There is a large cross-over between automated decision making and artificial intelligence, therefore they are merged together in a single document.
Good
- Less strain on workers
- Real-world application
- Medical diagnosis
- Pattern recognition
- Artificial intelligence
- Fraud detection
- i.e., what Stripe does; pattern recognition
- Anything in day-to-day life
- Traffic lights
- Recruitment
- Stock market bots
- Medical diagnosis
- Intended to be fair and unbiased
Bad
- Who takes responsibility for the decision?
- What if the fault for the decision is not due to them?
- What if the fault for the decision is due to atmospheric interference?
- Can be make unfair/biased by poor data
- Lack of trust
- Recent technological developments
- Transparency
- (AI only) Environmental effects
- Intensive to run - requires obscene amounts of energy
- Employment effects
- May replace human jobs - higher unemployment rates
- Access to technology
- Access to technology to run LLMs (large language models) in expensive and hard to come-by