The material benefit as such resides in the fact that (1) they can potentially save users a great deal of time (a very large portion of search engine queries are straight-up questions, which the articles in the result set sometimes answer, but usually only partially, and it still requires a good chunk of time to plough through them just to get that answer) and (2) high-potency propaganda of any kind (not specific to the hasbara project, though it seems to provide a shining example of such) promotes anxiety, paranoia, and increased susceptibility to psychosis. In fact to some extent that is its very purpose.
So anything that helps abate the pernicious blight of (2) is potentially quite helpful, both materially and spiritually.
Also, people are going to be using AI chatbots for this kind of research anyway. Rather than hoping the larger players will do the right thing here (or at least avoid doing the wrong thing), it seems quite prudent for independent organizations to pick up the task on their own, and start creating their own bots for these purposes.
And “saving users time while searching” is more likely to be accomplished by improving your query->result pipeline and not by improving a tangentially associated technology
Also, I didn't specifically say these bots would save users time "while searching", i.e. while using a regular search engine. Though I didn't dig too deeply into the T4P proposal (so I don't know what they're proposing), my own guess is that people would seek out these bots as an alternative to general-purpose search engines -- for example as a sidebar on their favorite news sites.
I think you're overthinking it. This is like saying LLMs will never be useful because they hallucinate. That's a known issue, and yet of course they have been proven to be often quite useful nonetheless.
What it comes down to is, how often do they hallucinate, what's the negative impact when they do (both of which can be measured) and very importantly: for whatever their measured performance is, how does it compare to the next best alternative that users have?
It's not like they're trying to build a model to design a nuclear reactor in one go. It's just a Q+A bot, whose performance can be easily measured by benchmarking it against the top 30 questions or so in a given subject area (probably accounting for 95 percent of all inputs). And the current alternative users have (search engines) is pretty darn mediocre.
BTW I'm actually not much of a fan of LLMs or chatbots, so I have nothing to "sell" you here. But this is my rough take, based on my generally quite skeptical attitude toward this technology.
Which does seem to suggest that, at the very least, it's an idea worth exploring.
The problem is not how often but how bad just one single error can be. My point was on controversial topic, where a single error can deal serious damage. Yes it must be error-free like for a nuclear reactor. Just imagine a Q&A chatbot answering questions on the subject of Israel and Palestine or something else really touchy, do you really think you can afford any error/hallucination ?
[0] : https://arxiv.org/abs/2409.05746 [1] : https://arxiv.org/abs/2401.11817
My point was on controversial topic, where a single error can deal serious damage.
Okay, but so can a single garbage article on a search engine result. I guess one shouldn't build search engines then (unless they can be held to the same standards as nuclear reactors), because do you think we can afford even single bad result? Just imagine what will happen, etc.