is taking a look at it. in tennis there can be some dull moments. i can only imagine ai commentary. i have not listened to much of it. i can only imagine it being quite repetitive. you have three, four shots and a rally and then what? you get the excitement and you get the inner knowledge of strengths of players and pundits know better. according to ibm, they say, as the match is going on, the ai according to ibm, they say, as the match is going on, the a! has a fixed set of phrases that it can use according to the noise it hears. so it is not spontaneous, it is picking phrases that would be routinely used. ., ., , ., ., phrases that would be routinely used. ., ., ., ., ~ used. the other reason i do not like it, it starts used. the other reason i do not like it, it starts somewhere. used. the other reason i do not like it, it starts somewhere. i
generating for this year. this is the first time that we have generated ai commentary for tennis. we did try this for golf for masters earlier this year but we have had to take our large language models and train them in the language of tennis, learn the game of tennis through the stats we are collecting, and generate this commentary for the highlight reels. tell and generate this commentary for the highlight reels- highlight reels. tell me a little more about highlight reels. tell me a little more about why highlight reels. tell me a little more about why you highlight reels. tell me a little more about why you are i highlight reels. tell me a little more about why you are doing| highlight reels. tell me a little i more about why you are doing it and how you are actually doing it. you referenced some of their work but tell me about how you actually do this. ~ . , tell me about how you actually do this. ~ ., , ,., ., , this. we are using something that is our enterprise this.
an unforced forehand error on match point. earlier, i spoke to kevin farrar, the head of sport partnerships, ibm uk the people developing the use of this technology. yes, we re very pleased with what we are generating for this year. this is the first time we have generated ai commentary for tennis. we did try this for golf, for the masters earlier in the year. but we have had to take our large language models and train them in the language of tennis, learn the game of tennis through the stats that we are collecting, and then generate this commentary for the highlight reels. tell me a little more. before we talk about the end results, tell me a little more about why you are doing it and how you are actually doing it. you referenced some of the work, but tell me a little more about how you actually do this. so we are using something called ibm watsonx. this is our enterprise ai and data platform. we are training something called a large language model