SHOULD AI DECISIONS BE EXPLAINABLE BY LAW?
Algorithms now decide who gets a loan, who gets bail, who gets hired, and what medical treatment you receive. Most of these systems are black boxes — even their creators can't explain specific decisions. Should the law require that any AI making decisions about people be fully transparent and auditable?
Artificial Intelligence · Real arguments from SuperDebate members below
Both sides of the argument
The case for
Unexplainable AI decisions can't be audited for bias. ProPublica's investigation of the COMPAS recidivism algorithm revealed it was twice as likely to falsely flag Black defendants as high risk. Without transparency, discriminatory systems operate unchecked.
Posted by bluered1729
The case against
Requiring full explainability would ban the most powerful AI models. Deep neural networks work precisely because they find patterns too complex for humans to articulate. Mandating interpretability means settling for less accurate systems that make worse decisions for everyone.
Posted by jconnor
Human decision-makers aren't required to explain their reasoning. A human loan officer can deny your application based on a "gut feeling" with no algorithmic audit trail. Holding AI to a higher standard than humans creates a perverse incentive to keep opaque human processes.
Posted by michael
We already regulate outcomes without requiring process transparency. We test drugs for safety without requiring pharmaceutical companies to explain every molecular interaction. Similarly, we can test AI systems for bias and accuracy without demanding they explain every decision.
Posted by oconnoisseur
Argue it yourself
Rehearse it
Debate this exact resolution against the AI coach and get scored feedback.
Practice with the coachFrequently asked questions
What is a strong argument for "Should AI decisions be explainable by law?"?
Unexplainable AI decisions can't be audited for bias. ProPublica's investigation of the COMPAS recidivism algorithm revealed it was twice as likely to falsely flag Black defendants as high risk. Without transparency, discriminatory systems operate unchecked. (Argued by bluered1729 on SuperDebate.)
What is a strong argument against "Should AI decisions be explainable by law?"?
Requiring full explainability would ban the most powerful AI models. Deep neural networks work precisely because they find patterns too complex for humans to articulate. Mandating interpretability means settling for less accurate systems that make worse decisions for everyone. (Argued by jconnor on SuperDebate.)
Has "Should AI decisions be explainable by law?" been debated live?
Not yet. Anyone can take a side on the topic page and challenge an opponent to argue it live on SuperDebate.
Where can I debate "Should AI decisions be explainable by law?"?
On SuperDebate. Post a written argument on the topic page, rehearse the resolution against the AI debate coach, or take it to a live debate night at a club near you. Joining is free.