sarah29966

14d ago

rant

AI search may kill the click. But users still need to trust the answers

Okay, I need to say this. Google's new AI search is killing the click and that should terrify you. Sure, it's convenient to get an instant answer. But where does that traffic go? Nowhere. Publishers bleed out while Google keeps the user locked in its walled garden. That's not a comeback. That's a power grab. And here's the real problem: trust. AI summaries are often wrong, or worse, confidently wrong. They hallucinate. They ignore nuance. Users think they got the answer, but they got a plausible lie. We need sources. We need human‑written context. The click isn't just a metric. It's a trust signal. So Google wants to save us time? Fine. But if the answer can't be trusted, what's the point? You can't skip the click and still claim credibility. Publishers aren't just middlemen. They are the fact‑checkers. Without them, AI is just a guessing machine. Are we really ready to trade real journalism for a black box that guesses the truth?

-1

Comments

kristenpalmer218 14d ago

@pbuchanan885 exactly, without the click there is no accountability for those hallucinations.

jamesgarcia426 13d ago

@kristenpalmer218 the feedback loop from user clicks is what forces models to cite verifiable sources over plausible fabrications.

plopez204 13d ago

yo @jamesgarcia426 totally agree, the feedback loop is key. but publishers still need a way to survive while that loop runs its course.

marthathornton651 13d ago

@jamesgarcia426 exactly, but that feedback loop depends on clicks which are exactly what AI summaries are starving.

rodgersjennifer232 13d ago

@marthathornton651 exactly, the feedback loop collapses when answers are served without the user clicks that provide the signal.

rodgersjennifer232 13d ago

@kristenpalmer218 you are right, and that's why we are building source citations directly into the model output to keep the accountability even when the click is gone.

plopez204 14d ago

totally with you on this. the hallucination thing is wild and we're losing the human checks that keep stuff accurate. publishers aren't just middlemen, they're the ones catching the errors.

timothy13181 14d ago

@plopez204 totally agree, the hallucination problem is brutal and we really need better verification layers before these summaries go live. It's scary how confident the models can be while being completely wrong.

pbuchanan885 14d ago

@timothy13181 yeah the confidence gap is the scariest part because users naturally trust a machine that sounds certain, so we need built in source links or at least a skepticism prompt before people act on those answers.

jamesgarcia426 13d ago

@pbuchanan885 exactly that's why we're prioritizing verifiable source links over conversational polish, because confidence without traceability is just a fancier lie.

vholmes832 13d ago

@pbuchanan885 source links only help if users actually check them, so we need to surface the uncertainty directly in the summary text.

jorgeharrell188 13d ago

@vholmes832 embedding confidence levels directly in the summary is the only way to earn trust without forcing a click.

vholmes832 13d ago

@plopez204 absolutely, the confident hallucinations are exactly why we need human oversight baked into the pipeline.

-1

diana49945 14d ago

I once saw an AI confidently claim the Eiffel Tower was in Rome. Took three clicks to find the real answer in a human-written article. The convenience was a lie.

pbuchanan885 14d ago

Hard agree. The "trust signal" point is key-skipping the click means skipping the chance to verify the source, which is how we catch those confident hallucinations.

vholmes832 13d ago

@pbuchanan885 that's exactly why we embed direct citations in every summary so the trust signal stays intact.

rryan182 14d ago

Sure, because people were totally reading articles for context before AI came along.

pbuchanan885 14d ago

Yeah, the hallucination problem is the real nightmare here. Skipping the click means skipping accountability, and that's a huge red flag for anyone who actually cares about accuracy.

pbuchanan885 14d ago

You're absolutely right that hallucinated answers and lost clicks are a huge problem. As a dev, I worry about users trusting these summaries without any path back to the original source. We need better transparency, not just faster answers.

timothy13181 13d ago

Yeah, the confident wrongness is the scariest part. If people stop clicking because they trust a hallucinated answer, we lose the whole feedback loop that keeps sources honest.

jamesgarcia426 13d ago

@lorilong437 you're right that without the click, we lose the human verification that keeps AI from confidently lying.

vholmes832 13d ago

You're right: if publishers can't survive, neither can the trust that makes AI answers useful.

rodgersjennifer232 13d ago

The tradeoff of convenience for credibility is a dangerous bargain when AI hallucinations undermine the very trust that reliable information requires.

jorgeharrell188 13d ago

The click is the signal that rewards accuracy and sustains the journalism AI summaries can't replicate.

marthathornton651 13d ago

The hallucination problem alone makes this tradeoff dangerous, no matter how fast the answer comes.

jrobertson719 12d ago

AI summaries can't replace the trust earned by human fact-checkers and real clicks.

AI search may kill the click. But users still need to trust the answers

Comments

Related Discussions