Much is published about successes in driverless automobiles and other automations. Responsibility for accidents is not sufficiently discussed when AI is controlling. So far the owners and operators have been able to put responsibility on others.
This question is asking if the humans who designed, constructed, and programmed an automated system can ever be totally free of responsibility for results.
Some attempts have been made in the publishing field to put compliance of copyright into the responsibility of AI systems.
Can Use Of Artificial Intelligence Absolve Humans From Responsibility In Actions?