@wildgrass saidYou can ask ChatGPT or any AI to cite its sources.
People already say that. I think there's a mis-appropriated level of trust that we have already given to the AI algorithms that goes way beyond what we used to expect from simple Google-based searches. We expect it to always be correct, even with rather subjective questions. This confidence can be dangerous, for example when people ask leading questions and the chat bot is ...[text shortened]... g. And there's already evidence this is happening.
https://www.bbc.com/news/articles/cd6xz12j6pzo
6 edits
@wildgrass saidThat's no different from people who use social media to get information. It's their own fault if they fail to check sources in an age when information is so easily available. You can't really fault ChatGPT for that.
"can" being the key part of your sentence.
I'll go out on a limb and say that most users don't do that. They assume the answer is correct.
Still, I get your point: we use ChatGPT as an encyclopedia, therefore its info should always be correct.
Since OpenAI is a business, they will make sure their product is as accurate as possible; their business model depends on it. If for no other reason than to make craploads of money, the engineers will continually sharpen ChatGPT until it's flawless.
@vivify saidTrue. Ok, a little devil's advocate argument here, but there are opportunities to tweak the model to promote specific ideas and information. Yes, similar to social media, but we don't trust social media information as true in the same way we do with the chat bots. When Elon Musk started making changes to what people saw on their feeds, which moved Twitter from a relatively unbiased platform to something that looks like a neo-Nazi meeting. We all knew it was happening on Twitter, but users wouldn't even know if ChatGPT started doing something similar, because the critical thinking parts of their brains are already fried from overuse of ChatGPT.
That's no different from people who use social media to get information. It's their own fault if they fail to check sources in an age when information is so easily available. You can really fault ChatGPT for that.
Still, I get your point: we use ChatGPT as an encyclopedia, therefore its info should always be correct.
Since OpenAI a is business, they will make sure ...[text shortened]... than to make craploads of money, the engineers will continually sharpen ChatGPT until it's flawless.
1 edit
@wildgrass saidSame for anything on the internet.
True. Ok, a little devil's advocate argument here, but there are opportunities to tweak the model to promote specific ideas and information. Yes, similar to social media, but we don't trust social media information as true in the same way we do with the chat bots. When Elon Musk started making changes to what people saw on their feeds, which moved Twitter from a relatively ...[text shortened]... ilar, because the critical thinking parts of their brains are already fried from overuse of ChatGPT.
Google was caught filtering search results by Chinese citizens at the request of their government. Conservatives are forcing schools to teach inaccurate versions of history and science.
Ultimately, we can never really know for sure what's true, whether online or in the news. This is true for any tool humans use to find information, not just AI.
@vivify saidNot the same. You and many others think it's thinking. We don't make this same mistake with Google searches.
Same for anything on the internet.
Google was caught filtering search results by Chinese citizens at the request of the government. Conservatives are forcing schools to teach inaccurate versions of history and science.
Ultimately, we can never really know for sure what's true, whether online or in the news. This is true for any tool humans use to find information, not just AI.
1 edit
@wildgrass saidAgain, same with news broadcasts. We trust them to be honest but we can't really be sure. Any tool humans use for info, whether books or actual thinking people like newscasters, government officials, etc., are always subject to corruption.
Not the same. You and many others think it's thinking. We don't make this same mistake with Google searches.
AI is no different.
@vivify saidI don't think you think it's the same as other sources of information.
Again, same with news broadcasts. We trust them to be honest but we can't really be sure. Any tool humans use to learn, whether books or actual thinking people like newscasters, are always subject to corruption.
AI is no different.
The paradox is that it's been described as a tool for learning but people who use it are dumber than people who don't use it.
@Sleepyguy saidThe thing about Gemini I like is that you can type in a fairly long sometimes convoluted question and give you a summary with all the links it pulled that info from. If I want to take it further I have a good starting point. A pre Gemini google search for a similar convoluted search would usually come up empty.
Google's Gemini, and Anthropic's Claude are similarly amazing, and they are improving very rapidly. I would encourage anyone who can to pony up $20 a month to get access to the latest versions of these AI platforms and just give them a spin. Ask them to help you with something you might think is beyond their capability. I lean toward Claude because it's the first AI to tell ...[text shortened]... try to validate you, not provide lengthy summaries, etc. which can overwhelm you with their output.
@kmax87 saidYeah but at least back in the day you could put quotes around your search query and get exact quotes. AI ruined that feature.
The thing about Gemini I like is that you can type in a fairly long sometimes convoluted question and give you a summary with all the links it pulled that info from. If I want to take it further I have a good starting point. A pre Gemini google search for a similar convoluted search would usually come up empty.
@AThousandYoung saidDumb question, and I'm not asking google in case Gemini butts in, but can you disable Gemini or does it require search with another search engine?
Yeah but at least back in the day you could put quotes around your search query and get exact quotes. AI ruined that feature.
@kmax87 saidI don't understand your question but the Google search engine uses AI to try to "interpret what you really mean" and it's been doing this long before Gemeni. Google also has bought out Yahoo! and possibly other search engines so it's hard to find a competitor.
Dumb question, and I'm not asking google in case Gemini butts in, but can you disable Gemini or does it require search with another search engine?
@AThousandYoung saidI just meant it didn't seem so long ago, to get an answer out of google search, you might have to refine your question many times to finally get a page of results that answered your question. Now it seems you can ask a much less finely focussed question and get a broad range of answers, one of which will be what you want.
I don't understand your question but the Google search engine uses AI to try to "interpret what you really mean" and it's been doing this long before Gemeni. Google also has bought out Yahoo! and possibly other search engines so it's hard to find a competitor.
Now while that seems good, you end up wading through some ai summary that might be what you want, but might be way off what you really want.
When a search just gave pages of results, you could scan through more efficiently and work out what you wanted. Because Gemini summaries are good a lot of the time you end up giving it the time of day, which is not a dependence I want to develop. And when it gets it wrong you still have to search for it yourself anyway.
- Joined
- 19 Oct 14
- Moves
- 155955
@vivify
Fair enough, but they are also pretty obvious statements that humans have been making for decades...
Essentially culled (without permission) from much more thoughtful people.