If an artificial intelligence system produces a recommendation or decision for a government entity or state officer that could impact the rights, duties, or privileges and immunities of a person, the recommendation or decision must be reviewed and is subject to rejection or modification by a human in an appropriate responsible position.
