John Deere to pay $99M in right-to-repair settlement

Why Read This from an HCI Perspective

This isn’t just a right-to-repair issue—it’s an HCI case about how much control users retain even after they buy a product. When a machine breaks, users don’t only need ‘repairs’; they need to understand what’s blocked, decide who to get help from, and determine when to stop the work. This connects directly to explainability, controllability/ability to intervene, and failure handling in AI services.

CIT's Commentary

The John Deere case clearly shows that even with strong technology, user experience can quickly fall apart if the interface is closed. The real problem isn’t simply permissions for parts or software—it’s whether, in a failure situation, users can read the system’s state and choose the next action. From this perspective, providing ‘diagnostic tools’ isn’t just adding features; it’s closer to restoring user control. Especially for safety-critical systems, the design hinges on how far users can intervene and what paths are available when something fails. Industry often frames this as a matter of regulation and maintenance costs, but research can treat it as a starting point for asking about state transparency, recoverability, and minimum requirements for human-in-the-loop design.

Questions to Consider While Reading

• Q.How can we verify whether providing repair tools actually leads to autonomous diagnosis and recovery in real-world settings—or whether it still only allows limited, predefined procedures?
• Q.What information structures and visual representations should an interface use to help users easily understand system status when something fails?
• Q.If we design intervention permissions for AI agents or autonomous systems that correspond to ‘right-to-repair,’ what level of user control should be the default?