Kovacs' Bert AI --The future of AI instruction?

[doctorhl]

Anyone use Kovac's Bert AI? It is a cell phone app that will biomechanically analyze any tennis stroke that you film and upload. It will then generate corrections and recommend drills using artificial intelligence. The AI is generating its feedback by scanning Marc Kovak's voluminous library collection of videos of stroke analysis to compare it with your uploaded video. My question is not based on the quality of Kovac's library or the software program, but rather the pros and cons of applying AI to tennis instruction. This instructional method has major implications for future sport coaching and teaching and has potential for misapplication of its analysis and recommendations in the student's hands.

[stotty]

I don’t think there is anything out there yet (at least last I looked) that can give accurate biomechanical feedback. Maybe glaring faults…but nothing too deep or granular. AI seems to fair much better when it comes to matchplay analysis than it does in technical analysis.

I imagine the only way to get accurate feedback would be to input 3D video of some sort because camera angles in 2D would be hard to interpret.

Someone sent me a link to SportsAI which I haven’t tried yet, but according to a colleague it’s not that great…yet.

I like Kovacs but find his 8 stage method a bit limiting because it can miss things; like forward entry, for example. I tend to stick with BG’s methodology. I understand it better and it works.