Bryson DeChambeau Uses Google Gemini to Fix Swing

Multiple outlets report that golfer Bryson DeChambeau used Google's AI assistant, Gemini, to analyze and troubleshoot swing mechanics after a frustrating late-night practice session at LIV Golf Korea. Per Fox Sports and Golf.com, DeChambeau described asking the AI about "physics principles" including "alpha torque and gamma torque" and said the discussions led him to reduce grip pressure and relax tension. He finished third at the event, one shot out of a playoff, after closing with a final-round 65, according to Sports Illustrated and Golfweek. Coverage ranges from straightforward reporting of the quotes to editorial reactions noting the novelty of a tour player consulting a consumer AI assistant in competition week.
What happened
Multiple outlets including Fox Sports, Golf.com, Sports Illustrated and Golfweek report that Bryson DeChambeau consulted Google's AI assistant Gemini to diagnose swing problems during LIV Golf Korea. Fox Sports quotes DeChambeau saying he spent "long hours on the range" and then "was talking to AI quite a bit last night trying to go through some different physics principles that makes the club turn over, having some alpha torque and gamma torque put in there." Golf.com and Sports Illustrated report he identified grip pressure and tension as likely contributors and said relaxing his grip helped him feel the club turn over during the final round. Sports Illustrated and Golfweek note he closed with a 5-under 65 to finish third, one shot out of a playoff.
Editorial analysis - technical context
Using a conversational large multimodal assistant such as Gemini as an on-the-fly diagnostics tool fits an emerging pattern where athletes and coaches adopt consumer AI to augment domain expertise. Industry-pattern observations: practitioners commonly use AI for rapid hypothesis generation, literature retrieval, or translating biomechanical descriptions into testable coaching cues. For example, a golfer can ask an assistant to enumerate torque concepts, sequencing cues, or simple drills, then test those cues on the practice range. This usage does not replace biomechanical measurement but can accelerate the ideation cycle between observation and intervention.
Context and significance
Editorial analysis: The story is notable because it publicizes a high-profile athlete using a mainstream generative AI assistant during competition week. For AI/ML practitioners, this illustrates broader consumer adoption pathways outside enterprise and research settings. Observed patterns in similar consumer deployments show value comes from lowering the barrier to domain-specific reasoning, while risks include overreliance on surface explanations, hallucinated physics, or advice lacking sensor-backed validation. Reporting also captures mixed public reaction, from curiosity to ridicule, which mirrors prior debates about AI in public-facing roles.
What to watch
Editorial analysis: Observers should track:
- •whether athletes pair conversational assistants with objective sensor data (motion capture, launch monitors) to validate AI-suggested cues
- •how teams and coaches integrate AI into established coaching workflows
- •any vendor moves to productize sports-specific advisory features or to advertise athlete endorsements. Media coverage may also influence public perception of AI as a legitimate training aid versus a gimmick
Limitations
What happened: None of the cited reports include technical logs or the AI prompt history, and DeChambeau has not published the Gemini interactions. Industry observers therefore cannot verify which model behaviors produced the cue that DeChambeau described.
Scoring Rationale
The story documents a notable consumer application of generative AI by a high-profile athlete, which is interesting to practitioners tracking adoption patterns. It does not report a technical advance or new product release, so its direct impact on ML work is limited.
Practice interview problems based on real data
1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems

