Historians reading newspapers learn what a small part of society thought, and hopefully read them in the context of as much other information from the time. But maybe the newspaper made something that hardly anyone really cared about seem like the main opinion, or vice versa - it happens now days, likely more back then.
If anything social media might make things easier for future historians, with literally billions of opinions written by everyone from homeless people to billionaire CEOs, all within large datasets to be analysed.
More importantly though is the amount of semi-political & marketing-related polling that goes on. By semi-political I mean things politicians want to know, rather than polling about politics - people regularly get polled on anything from transport ideas to favourite songs. And I imagine that's a huge trove of data compared to what there is around public opinion even decades ago, yet alone centuries.
Ubik, by Philip K. Dick
Pre-internet everyone consumed the same media, TV, newspapers, radio, magazines. You had to read older books or listen to smaller, probably local artists to get a different viewpoint.
Now no one shares the same information feed, no two Tiktok feeds, Twitter feeds, Facebook/instagram homepages are the same.
It's making for interesting societal changes. How will future historians be able to judge the intentions of society of they can't view our individual feeds, previously you could refer to news reports, but now it's just lost in the day to day feeds of billions.