4 ms·
I agree that article's headline is pretty factual nor are the facts cherry-picked. But "some reporting is biased" is heavily understating the problem. And I hyp
by zaphar 1y ago
I agree that article's headline is pretty factual nor are the facts cherry-picked. But "some reporting is biased" is heavily understating the problem. And I hypothesize that the reason there is a leap to "This is biased" today on reporting is because the news media organizations have participated in the cherry-picked facts case to enough of a degree that trust has eroded.
Which is perhaps why organizations are searching for ways to not appear biased and begin to restore that trust. I think the reason the LLM approach is likely to fail is because it is unable to detect the "missing facts" case and can only really advise about sentiment and phrasing. Which is not I think the actual problem that needs fixing.
- JoshTriplett 1y agoI agree that an AI is unlikely to help. But also, I doubt that the primary reason is "restoring trust". I think the primary reason for many of them is that some of their readers react strongly to things they don't want to hear, and they're afraid of losing customers, so they're watering down their reporting to avoid being inconveniently right. Or, even less charitably, management and employees have different politics, and management are the ones who find the articles inconvenient. To be clear, there are absolutely biased news sources out there. For many of them, the bias is the point, and they have no particular desire to "restore trust" because they're already trusted by people who only want to read things supportive of their party. But a politician who finds the truth inconvenient will decry everything accurate as biased.