Such actions have ripple-effects.
Example for weather-amateurs: historically NOAA is linked to CWOP at one side and NWS at another side.
CWOP injects meteodata from private weatherstations all around the globe.
Likewise NWS also gets meteodata from various other sources like Weatherflow (
see WF's webpages!), which in turn relies on free-of-charge input from linked private weatherstations (especially a large amount in the US).
NWS processes the received meteodata and translates to 'consumable data'.
AI might help in data-processing, but a model is as wise as the quality of it's inputs and the setup&maintenance of the model.
IMHO NOAA and NWS have useful responsibilities for that aspect.
Also because of the nature of weather prediction most often still human knowledge, experience and view is required to ultimately make sense of interpretations .......
Apparently the US-government assumes that companies
like IBM WeatherChannel (incl. WUnderground) seamlessly and without cost-increase take over all services&responsibilities from NOAA and NWS.
Are they also expecting that those private weatherstations continue providing their data free of charge, if the resulting output would be commercially exploited?