By Lee Flanagan
Weekly Roundup · 2026-08-17 to 2026-08-21
A candidate walks out of a sixth interview round and the video goes viral. Eightfold, Meta and IBM face lawsuits alleging AI hiring bias. The Guardian publishes a takedown of AI interview bots. Three stories, one throughline: when candidates cannot see how you reached a decision, you own the risk.
AI Bias Lawsuits Hinge on Explainability, Not Just Fairness
The legal exposure is not that the algorithm got it wrong. It is that no one inside the company can explain why it scored a candidate the way it did. When hiring teams cannot articulate the reasoning behind a rejection, they cannot defend it in court or to the candidate. Explainability gaps exist long before AI enters the process, but automation makes the liability visible and actionable.
Interview Round Count Is Not the Problem. Redundancy Is.
A viral TikTok walkout reignited the debate over how many interview rounds are too many, but the real friction is not volume. It is asking the same questions, testing the same competencies or requiring the same case study three times over. Candidates tolerate rigour when each stage adds new evidence. They walk when repetition signals disorganisation or a lack of internal alignment on what the role actually requires.
The Backlash Against AI Interviews Misses the Structural Flaw
The Guardian blamed Zoom and AI bots for broken hiring, but the real problem is unstructured, low-transparency interviewing, regardless of who conducts it. A human interviewer asking vague questions with no scoring rubric creates the same opacity and inconsistency as a poorly designed bot. The technology is a convenient scapegoat for a process flaw that predates automation by decades.
The common thread is not AI, interview length or candidate expectations. It is opacity. If you cannot explain to a candidate, a hiring manager or a regulator why you made the decision you made, you are carrying a liability that scales with every rejection you issue. The question is not whether your process is fair. It is whether you can prove it.
