Social high school sports data provider Scorestream has completed an eight-month pilot program with Abundat’s Lede Ai to write automated game stories using as few as 10 data points.
The beta test stories were published by the Richland Source newspaper in Mansfield, Ohio, the same town where Abundat is headquartered. The Source began producing AI-generated stories in August 2018. In the first six months of the trial it posted 18,000 high school game stories from across Ohio using Scorestream data and Lede Ai writing.
A Lede Ai white paper summarizing the project indicated that there were no reported inaccuracies. The whole process was automated so that each story was written and published without any staff intervention. Lede Ai is also active in seven other closed-beta trials. Scorestream was recently issued a patent for its crowdsourcing data technology.
SportTechie Takeaway
This Scorestream-Lede Ai collaboration provides a new use case for AI-written news stories. The Associated Press previously partnered with Automated Insights to create computer-generated minor league baseball recaps and NCAA men’s basketball previews, which both carry a reasonably high profile and wide reader interest. (ScoreStream has already partnered with the AP but, so far, only on scores.) Covering high school sports is an ultimate example of local journalism, and coverage has waned in locations where there aren’t the reporting resources to staff events. Lede Ai usually charged $4 per story created, and appears to be a scalable, budget-friendly solution.