An AI algorithm is set up by a sponsor, and is applied to subjects. When studying the risks of AI, keep in mind that the risks to a sponsor are entirely different from the risks to a subject.
The sponsor profits from the work done by that AI.
A sponsor is typically a corporation or government entity that deploys an AI.
If the algorithm is not performing well, the sponsor can change it. The sponsor can update the algorithm’s decisions if they don’t serve the sponsor’s needs.
A sponsor is typically isolated from the effects of many false positives or false negatives.
The sponsor benefits from transaction volume - the more transactions, the better, even if some of those transactions are erroneous. Compensation comes from quantity, not quality of transactions.
The subject faces the consequences of the work done by that AI.
A subject is typically a thing or a person whose life is affected by an AI’s decisions.
If the algorithm is not performing well, the subjects cannot change it. They must live with the algorithm’s decisions whether they like it or not.
The subject is directly affected by single decision.
Algorithms can determine who gets a bank loan, who gets an organ transplant, or who gets parole. A single false positive or false negative can be harmful, even fatal to a subject. It can deprive the subject of their job, their home, their health, or their freedom.
Capitalism is founded on the premise that if a customer doesn’t like a service, they can shop around for a better one. With AI algorithms that make decisions about people, the subjects cannot “shop around” for a better algorithm. Sponsors typically enjoy a monopoly in markets where they deploy their AIs.
Typically an AI’s behavior is not independently verified, so the sponsor has the absolute and final say over whether that behavior is correct.