Grok Turned a Basketball Metaphor into a Defamation Story About Klay Thompson

What happened

In April 2024, Klay Thompson had an "all-time rough shooting" night for the Golden State Warriors, hitting none of his shots in his final game with the team. The morning after, X's Grok AI chatbot surfaced a post titled "Klay Thompson Accused in Bizarre Brick-Vandalism Spree," describing Thompson going on a criminal spree of vandalizing homes with bricks. Thompson had committed no such act. No such accusation existed anywhere. Grok had invented the whole thing from a basketball term.

The mechanism was embarrassingly literal. In basketball, a missed shot is called a brick. After Thompson's historically bad performance, sports commentary was saturated with the word. Grok, processing that coverage as raw text without understanding the domain-specific meaning, appears to have concluded that the stories described an actual person throwing actual bricks at actual houses. It then assembled that reading into a post styled as a news report. The output was confident, complete, and entirely wrong.

X ran Grok's post with a disclaimer in small type beneath the content: "Grok is an early feature and can make mistakes. Verify its outputs." The qualifier did nothing to arrest the story's spread. The headline framing was specific enough to read as reported fact, the subject was a recognizable public figure, and nothing in the platform's presentation distinguished a Grok-generated assertion from a link to an actual news article. The disclaimer was placed after the damage, not before it.

The incident landed in an ongoing legal conversation about whether chatbot operators bear liability for defamatory outputs. Commentators noted that Grok appeared unusually susceptible to misinterpreting figurative language, and described the system as vulnerable to producing misinformation from ambiguous inputs. The Thompson case was not isolated. Around the same period, similar complaints had surfaced about other chatbots generating false criminal accusations against named individuals, including one in which a rival system fabricated a fraud allegation against a radio host in connection with a lawsuit he had never been part of.

The accountability gap here is not subtle. A system that generates named-individual allegations and publishes them at scale, with no human review step between inference and distribution, creates defamation risk that a disclaimer cannot patch. The missing piece is a provable record of what a system did: how it reached an assertion about a specific person, whether any check was applied before the content reached an audience, and who was responsible for that decision. Without that record, a bad shooting night becomes a criminal accusation, and the only correction mechanism is public ridicule after the fact.

Reported impact

Affected parties

Not publicly disclosed

Harm type

Not publicly disclosed

Scale

Not publicly disclosed

Financial impact

Not publicly disclosed

Regulatory action

Not publicly disclosed

Classification

Organization

Not publicly disclosed

AI system

Not publicly disclosed

Industry

Not publicly disclosed

Country

Not publicly disclosed

Provider

Not publicly disclosed

Incident type

Not publicly disclosed

Relevant governance controls

Governance control mapping is not available for this record.

• No controls mappedNot publicly disclosed

Control mapping is analytical. It does not state that any control would have prevented the incident.

Sources and evidence

This record was researched and written by the Index. The event is also catalogued in the following database, which is listed for cross-reference.