Automation bias is the tendency for humans to trust that processes which are automated are also more reliable. Alternatively, humans tend to trust manual processes as less reliable.
While automated processes can be theoretically more accurate than manual processes, it’s also possible that the automation is not implemented to that level of performance.
Errors stemming from automation bias tend to happen when a user trusts an automated system without question, even if the results would appear untrustworthy given close inpsection and critical thought.
Examples of Automation Bias
A self-driving car making an incorrect turn, or attempting to merge into a lane while another car is in the way.
Automation bias could be observed by a user trusting the automated driving software and thus not paying attention to the decisions made by the car. While self-driving can be theoretically better than manual driving, curren systems often make many mistakes that a human otherwise would not make.
A GPS suggesting a route to a destination that is slower than the optimal route, due to a new road recently being constructed.
Automated software only understands what it's been programmed to do. In this case, the GPS may not be aware of the optimal route, if recent construction has altered the available routes.