As far as they are concerned it's just automating in a superior data mining fashion a profile on you. They aren't doing it for just for you. There's always a catch.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Google Now Testing Generative AI for Search
- Thread starter MacRumors
- Start date
- Sort by reaction score
For one idea how this “works”, I never use the feature to “save password” because it’s one way to enable tracking. After logging in to try Google’s AI music toy, it doesn’t just log you into that session, it keeps you logged in when you leave the page. I’d imagine all the AI players are pulling in a good quantity of data even for folks “passing by” as they drop a ton of cookies.As far as they are concerned it's just automating in a superior data mining fashion a profile on you. They aren't doing it for just for you. There's always a catch.
LOL ... Google's AI is simply to point you to more paid website ad listings first and foremost then more relevant sites down the results page - just laid out nicely so you cannot tell which is which
Likewise ... but even BEFORE Google's AI use, and before AndroidOS with Google's services (not the core AndroidOS it self, as must at least), the trust issues with Google for me initially stemmed from paid ads which pointed to sites at the top of search results.
Whom said Google's AI was made or would do anything to alleviate trust issues with Google?
Those using AndroidOS with Google's services sold their souls for the cheapest entry to what they wanted over a decade ago.
Microsoft already found researchers using non-user data input injection techniques to evoke and change Bings LLM and made rapid adjustments and continually working to find such rogue sites and security concerns. This makes me feel a LOT better using their M365 office services when CoPilot (using ChatGPT LLM + internal endeavours) next month.
THAT'll be HUGE!!!!
This is the real danger of “AI”: People believing that LLMs like ChatGPT provide credible knowledge. LLMs are basically opaque, “fuzzy googles” as they take the same input (in the case of ChatGPT curated website and article scraping) but provide a non-referenceable average of the most common token (text bit) relations as the output. Traditional search engines use a form of “AI” to “understand” the search query and relate this to an index, and googles novel approach will most likely add “understanding” the scraped content to the formula in order to provide better results with qualified summaries rather than the current simple excerpts.
my long-time problem with google search was that it already didn't really searched for the terms i gave, but extended them using some hideously stupid ways i couldn't fully control. and there are the ads.
and now generative AI. thanks, but no.
some say, if you build a thing in a way stupid people will be able to use it, in a while inevitably mostly stupid people will use it.
and now generative AI. thanks, but no.
some say, if you build a thing in a way stupid people will be able to use it, in a while inevitably mostly stupid people will use it.
Google search is pretty bad these days. It's mostly just ads.
Soulless corporation creates soulless design from soulless AI. -Dr. Ian Malcom
Of all the technological advances we've seen, this AI race makes me the most uncomfortable. I guess I should have read all the terms and conditions I've been agreeing to all these years
I have no idea what you report on, but most journalists scrape Twitter for their stories/news and don't really fact check or present any of their own work when they write an article or are a talking head on TV anyway. It's why the the legacy media isn't trusted or really relevant anymore.
It's thanks to the legacy media (BBC, CNN, Sky/Fox News, Daily Mail etc) that fake news & bias is as widespread as it is anyway.
Maybe journalists are going to have to actually do a little work before presenting a story to the public by checking with sources before rushing out with something that is later proven false/fake/misleading/irrelevant.
IMHO I don’t see Apple at all interested in competing using a search engine even if it is AI based. The only thing they need to provide is a choice of what search engine is used by default within Safari settings. They offer several currently.
I can't quite believe how good Bing is now thanks to GPT-4, particularly the app. And baking Bing with GPT-4 into Edge and Windows 11 just makes sense and is highly beneficial to the user.
And yes, Safari is looking quite laughable compared to Chrome and Edge. But didn't it always?
I can’t comment on what specific reporters do but yes some scramble to be the first to get something to air/online without proper fact-checking. Seems to be younger, inexperienced reporters - perhaps some just trying to make a name for themselves or climb the ladder.
I’m old school - I’d rather be right than first. It annoys my bosses many times but I always ask would you rather a lawsuit instead of a few extra clicks?
If you have trust issues with google, you should be also worry about Apple. Their updated privacy policy is nothing but a thin vail used to increase their profits. what is does is block other companies from collecting your data, but apple still does and therefore gets all the data they want.
Apple is sneaking around its own privacy policy — and will regret it
Apple has a complicated relationship with privacy. It loves to tout its efforts, especially as a differentiator with Google. But actually delivering privacy? That’s a different story.
www.computerworld.com
Apple's Privacy Promises Are Undermined by Its App Store Rules
A blog post on the loopholes in app tracking policies like Apple's ATT and how new legislation can help solve these issues.
publicknowledge.org
Google is playing catch up to Bing for a change.
I don’t like all this generative AI popping up everywhere. Social media algorithms already have caused untold damage to society.
I don’t like all this generative AI popping up everywhere. Social media algorithms already have caused untold damage to society.
I hate this. Google search results are increasingly garbage, as the algorithm tries to directly come up with an answer to my questions, instead of directing me to where I can find the correct answer myself. AI will only generate another large source of false answers.
I think this AI search will be a fad for a while and then people will hanker for other websites again.
Wikipedia has most of the answers on it, but nobody goes there to get all their information. Same with search. Who wants to read answers from Google every single time?
Wikipedia has most of the answers on it, but nobody goes there to get all their information. Same with search. Who wants to read answers from Google every single time?
Yep. AI itself is not the thing we must fear. What we must fear is humanity itself, and what the average human will believe AI is and what jobs the average human is willing to let AI take on.
In the US, we have legislation on the table to ensure that nuclear weapons cannot be launched by an AI. If you've ever seen WarGames, you know exactly what this scenario could result in. But the fact that this is even in active discussion makes me worry - what else have people proposed that we should let AI do? Maybe this is pre-emptive (maybe even because of WarGames itself?), but I would not be at all surprised if somewhere in the military, someone suggested doing this, and (thankfully, for now) the rest of the room said "are you f--king crazy???"...
Professors will feed chunks of paper into ChatGPT and ask it "did you write this?" and it almost always will say "Yes, I likely wrote that text". Already seen plenty of stories where a professor threatened to fail more than half the class because of this.
We also must fear our own fallibility, the fact that we as humans can be so easily tricked by AI (which is a genuine fear of mine with respect to elections in the future). Simultaneously, we have to fear the level of trust we already appear to be willing to put into AI.
And now this expands it to anyone who has put time and effort into making a great website about birds or astronomy or whatever. Or contributed to wikipedia. Or written well-researched and thought-out forum posts/replies.
This thing will scrape all that data and repackage it into "AI generated by Google".