I'm pretty sure they have partnerships with the major stock and currency exchanges, pro sports leagues, and weather.com to get that data.

EDIT: added some sources (sorry, that's the best I could dig up in 5 minutes)

[1] https://www.google.com/googlefinance/disclaimer/

[2] https://searchengineland.com/google-now-with-real-time-nhl-h...

[3] https://bits.blogs.nytimes.com/2015/03/31/ibm-scores-a-weath...

US weather data is available for free from NOAA as long as the API isn't abused. Everybody else is just reselling the same data anyway.

A good comparison here is WolframAlpha, which can also answer all those queries (I was a bit surprised that the Texas Rangers example worked, but it does [1]). Most of their data comes from a curated list of primary sources, not Wikipedia or random websites.

1: https://www.wolframalpha.com/input/?i=texas+rangers+next+gam...

Curated might be better, Google has been wrong in their answers in the past.

Not only are the sources of WolframAlpha curated, they even go the extra step to make sure it's correct. From their FAQ:

"How is Wolfram|Alpha's data checked?

We use a portfolio of automated and manual methods, including statistics, visualization, source cross-checking and expert review. With trillions of pieces of data, it's inevitable that there are still errors out there.

[...]

How is real-time data curated?

Wolfram|Alpha effectively checks real-time data (such as weather, earthquakes, market prices, etc.) against built-in criteria and models. If an unexpected deviation is found, Wolfram|Alpha will normally indicate it, for example by showing lines as dashed."

https://www.wolframalpha.com/faqs/#data-in-wolfram-alpha

I think the best one I've seen was when Google's natural language processing mistook the subject of a sentence. Rather than saying that the Mariner 10 space probe used a gravity assist on Venus to reach Mercury, it instead said that Mercury received a gravity assist from Venus.

Or maybe Google was getting its astronomy from a site that believes Immanuel Velikovsky's theories... [1]

[1] https://en.wikipedia.org/wiki/Worlds_in_Collision

I think they generally pay for it. It came up with song lyrics a couple of weeks ago, where Genius accused Google of scraping lyrics from Genius, but as it turns out Google payed a third party vendor LyricFind for them, and LyricFind pays the copyright holder, so if anything it seems Genius was more in the wrong...

https://www.theverge.com/2019/6/18/18684211/google-song-lyri...

How was Genius in the wrong? They were not paid for the service, and the data was coming from them. It is possible LyricFind stole the lyrics from Genius, but then Genius is still not in the wrong to sue them for scraping their site.

So LyricFind says they licensed the lyrics from the publisher and then verifies against other lyrics websites. Somehow, they ended up with a copy of the lyrics as posted by Genius.

Genius doesn't own the copyright to the lyrics and screwing with the punctuation doesn't create a new work, so I'm not sure they have much of a case against anybody. Maybe they could come up with a ToS violation or CFAA case against LyricFind?

How did that site get the sports schedule? Did they just copy it from somewhere else?

And that's part of the question: If Google's getting all it's data from first party sources, say, scraping it from the team's website, it may not have harmed anyone there. But if it's scraping it from a site which aggregated and normalized the information into an easily parse-able format (hence, providing value, of which Google took advantage), then they must be compensated for it.

Presumably the site being scraped is getting some value from being included in Google's index. If they aren't, they can always opt out.

First you have to find a way to prove it’s your data which isn’t trivial.

Genius had to invent a watermarking system for texts: https://betanews.com/2019/06/17/google-genius-com-lyrics/

You don't have to prove anything to opt out of indexing.

And in the lyrics case, it wouldn't have helped because Google wasn't scraping the lyrics. Google was paying a third party that copied the Genius site.

So it’s a multilayered problem, and content owners are on the losing side.

If it's getting it from a site that's adding no value (e.g. just copying and pasting another site), then there's probably nobody to compensate either.