AI-Assisted Developer

AI coding assistants are now part of many developers' daily workflow, producing statistically average code trained on a corpus full of the Lucky 13 (see AI-02).

Using an assistant does not transfer responsibility for the output. The developer who commits the code, and the manufacturer that ships it, remain accountable for the product defects it contains.

Most manufacturers have not changed how they review code to reflect that. The author changed; the review did not.

Discussion prompts

• What in your code review process changes when the author is a model rather than a person?
• When AI-generated code ships a defect to customers, where does accountability land in your organization today?
• If your organization wrote its first policy on AI-generated code, what would the opening line say?