Google wants to target you – yes, YOU – with AI-generated ads Google plans to roll out generative AI tools that can automatically create online advertising campaigns personalized to users' search queries. The Alphabet-owned ad giant is ramping up efforts to inject more AI into its core business areas – such as internet search and digital adverts – as it faces challenges to its dominance …
would this not count as personal data and thus have to be regulated by the GDPR ?
Somehow Google seems to consider itself as not having to bother with this. It is not just using the search engine that profiles us but also things like google analytics that is on most web sites without asking for permission (including el-Reg - tsk, tsk).
However the UK government seems bent on watering data protection laws down and anyway the ICO is a bit of a chocolate teapot.
El Reg works well enough for me with Google Analytics (is that really a word?) blocked.
As do I (well, map it to 127.1.1.1). Maybe an El-Reg hack would like to add a comment below saying why it needs Javascript from: www.googletagmanager.com www.google-analytics.com pagead2.googlesyndication.com
I'm sure the Hangzhou Sage Chemical Company would have some.
(Yay! An excuse to link to the utterly hilarious, and not recently seen around here, Things I Won't Work With blog :) )
Or will they somehow use AI in the ad blocker arms race?
Probably. A few days ago, I fired up YT in Firefox and it disabled all my extensions. I don't recall giving AlphaGoo authorisation to disable security services on my PC, so I'm not entirely convinced this is legal. But then AlphaGoo doesn't care about what's ethical, moral or legal.
It also doesn't care about it's advertisers. If it did, there would be a way to alter recommedations with a prominent FOAD button to indicate that you have zero interest in their recommendations. I'd really love to know how AlphaGoo think that works because it never seems to have any idea about what I'm actually interested in.
Are you sure it was in Firefox?
Absolutely. It's the only browser I really use. The extensions had been working in other tabs, but as soon as I went to YT, they were gone. All were set to disabled, and I was prompted with an offer to make the pain go away by paying for YT 'Red'. Strange thing was I'd closed Firefox on a vid I was planning to watch rather than it's useless 'recommedations' page. It played that vid for <1s then cut to telling me I had 1:30s or so of ads, starting with a dubious looking one for some finance thing.
But on this point.. Anyone know any good app logging tools so if this happens again, I might be able to capture what happened? I also did the usual malware scans to see if anything else may have been to blame, but those all came back clean.
I'm skeptical. If this happened, then we should all smash our devices and run for the hills in tinfoil hats because all bets are off.
I don't question your personal experience, I just question the fundamental cause. If YT – or indeed any website – can mess with a browser's extensions in that way, then something is seriously wrong with the browser. Mayhap you should check for malware?
I don't question your personal experience, I just question the fundamental cause. If YT – or indeed any website – can mess with a browser's extensions in that way, then something is seriously wrong with the browser. Mayhap you should check for malware?
Agreed. I mentioned it because it was unexpected, and I take precautions. I didn't think it was something a website was (or should be) able to do. Also why I'm curious about any app or event logging that might help, if it happens again. Also malware scanning was the first thing I did after it happened.
Websites can't disable Firefox extensions. If they could then UBlock, AdBlock, etc would all be disabled on every website.
Something else was going on and you just happened to have a YT page open.
If you don't agree show me the Firefox API that allows a website to disable extensions. (don't waste your time, there isn't one)
Websites can't disable Firefox extensions. If they could then UBlock, AdBlock, etc would all be disabled on every website.
And yet something did. Hence why I mentioned it..
If you don't agree show me the Firefox API that allows a website to disable extensions. (don't waste your time, there isn't one)
Again why I mentioned it, and asked if anyone knew how to log for stuff like this. Sequence was something like firing up the PC, firing up Firefox and doing my usual newsround to see what happened overnight. Extensions were running because news sites tend to run lots of ads. Then switched to YT tab, and extensions disabled. So then a malware check, ie virus scans, fired up wireshark to look for dubious connections etc.
But even if it were malware, it would have been a bit odd. Why would malware modify a browser in such an obvious way? Most malware would run via it's own C&C sessions, or a browser hijack would attempt to redirect traffic. Anyone even slightly security conscious would have been aware something happened when they were spammed with ads again. Plus coming shortly after Alphagoo said they were going to start blocking ad blockers, and only seemed to trigger when I landed on YT.. it's natural to be a bit suspicious.
(And it's not like 'undocumented' or broken APIs have enabled malware in the past..)
while I haven't seen my extensions disabled, for the last... month or so, I have noticed that when I open a link to youtube, a new tab opens allright, shows the content for half a second or so, then goes black. Like, ALL black (that' propa blocking!). I need to refresh it and then it loads the content (sans ads of course). It's a tad annoying and it might be either accidental, or accidental-on-purpose fiddling by google, cause every little helps, you know...
That said, I've noticed that since I see no ads on any sites at all, I've become oversensitive to them whenever I have to use other people's computers, it seriously annoys me to see ads ads those ads noisy, ads colour-ads-vibrant ads before ads the ads start ads of ads every ads fucking ads video ads ads ads. But then again, it might be my age.
Or will they somehow use AI in the ad blocker arms race?
That would only be required if they start inserting ads in the bodies of stories. Like you are reading a Reg article about Apple's upcoming Macs and in the middle of the text you see a non sequitur paragraph touting a Dell PC. Then I guess you'd need AI to alter the article text to cut out the ad. But it sounds like the ads would be the same ads we get today only supposedly better targeted "because AI".
The role I see at home AI performing someday would be curation. It learns what sorts of stuff I like to read so it goes to The Register and other sites I visit on a daily basis, along with others it decides to check itself and finds articles that interest me and provides summaries so I can decide which ones I'm interested in enough to read the whole thing. I'd be able to consume many more news sites in the same amount of time, and discover stuff that I otherwise might have missed thanks to my AI knowing my preferences and maybe finding a random article about one of my interests in a place I would never have found on my own.
Obviously I would only use something I can run locally, like on the NPU in my iPhone, and not in the cloud where an ad company like Google can collect my personal information and tilt my curation towards whatever sites advertise the most with them.
"Google's Performance Max tools will analyze a supplied landing page to generate appropriate headlines, text descriptions, and images to use in Google-served web ads"
The essence of good advertising has always been creating a message that strikes an emotional chord by saying or implying much more than just headlines and descriptions can offer. The proposed approach is highly unlikely to be able to produce anything but the most banal and ineffectual advertising. But of course Goooooooogle won't care whether it's effective as they get paid even if it's not and it's pretty hard for the advertiser to pin them down. The well documented poor return for advertisers on the current auction based ad placement already demonstrates this. The problem of course is that nobody dare not advertise these days, so lousy is safer than none despite the cost.
My complaint about adverts is the proliferation of the bloody things.
As to those generated by AI, I'm not bothered. Companies that have my information and know my entire purchase history make such a piss poor attempt at serving up items of interest (oh, you just bought a shaver, we think you'd be interested in a new shaver... or here's a money off coupon for nappies despite having exactly zero history of anything to do with babies) that AI can hardly do a worse job.
I agree - look on any "newspaper" website these days - and its full of every ad, banner, popup and flying blo*dy toaster you can think of, usually all at the same time!
What next "hit the monkey to win a free holiday"?
Really early 1990's advertising practices.
Advertisers dont want it - I sure as fook dont want it - I would much rather Google or whoever knew that I live in Scotland, like Man UNited football team and eat fooking KFC - to actually show me ads for somethiing id be intersted in (whats that - 10% off my next KFC? Thank you very much!).
And that, my friends, is what happens when you dont allow browsers, Google, anyone-with-any-intelligenmce, to learn what it is you might actually be intersted in...
But no, "Personal Information, maaaaan" - 'cos BBC...
If you habitually eat chicken pieces and this is a known known, then the chicken flinger isn't going to be interested in you because they already have you. Thus, no 10% coupon for chicken bits, but numerous adverts from everybody else trying to tempt you away from your avian goodness.
Exactly. The expected return from advertising to a consumer something they already frequently buy is minimal, unless you think you can tempt them away from a competitor. Well, there's the theory about "brand prominence", which is what supposedly justifies the excessive advertising of things like Coca-Cola and F-150s here in the US; but I'm dubious about its value, frankly.
Yes, I have all those great extensions, and a PIHole on my home network.... so, even if they get my search history, they will be very disappointed. Most of my searches these days are for "what else did this actor star in?" "What other bands did this musician play in?" "What are the definitions of this word?"
but wouldn't it be nice to introduce a 'diamond' subscription package where you, The Ultimate Privilaged get to see all the latest and greatest junk 'up to' 24 hours before the plebs on 'platinum' package, who get a headstart of 24 hours on the bottom feeders on 'gold' package? Every little helps to increase our bottom line, never mind the feeders...
To serve ads about skin lotions when you search for the symptoms they claim to cure? I thought the reason Google was making money on search ads is that if you are e.g. Palmolive you can bid on search terms like "dry skin" "dry hands" "chapped skin" "cracked skin" or whatever and when those terms are searched for you see those ads?
So Google wants to stick an AI in between people doing the searching and companies with products to sell related to those search terms and think they can do a better job? Let's say the phrase "Joe Biden really chaps my ass" becomes a meme for some reason so a lot of people search on it. The AI is going to learn from that, and next thing you know people searching on "Joe Biden" get an ad for Chapstick or Palmolive - and I'm sure both those companies would be really happy about having their ads delivered in completely useless ways instead of the old way where they could choose the search terms themselves!
Keep drinking the kool aid, and ignoring that Trump was buddies with Epstein, that Matt Gaetz was having sex with teenagers and so forth. Unfortunately predatory behavior is a "both sides" problem, except for idiots who think that kids reading a book about being gay will make them gay. Funny idea for the people who always say "guns don't kill people".
I never got the concept of user consent to "relevant" advertising. Turns out I was holding it wrong.
"Google AI will start learning about your brand to populate your campaign with text and other relevant assets."
Category mistake on my part. They never meant relevant to me.
I'm shocked.
"A Google search for "skin care for dry sensitive skin," for example, would display ads from skincare brands with headlines generated by AI to match the user's queries."
Well that's completely useless to me.
There was a time when Google delivered search results that were, um, relevant to the search terms. In the unlikely event that I might search for "skin care for dry sensitive skin" then I would hope to get some tips on, I dunno, diet, lifestyle choices, home remedies, just about anything other than whatever some megacorp maximally bid for my search terms. When did "search results" become the same thing as "paid advertising"?
Oddly enough I haven't used Google for several years.
-A.
After seeing what AI has done so far, the ads are gonna be a hoot! I'll certainly enjoy watching the fallout when the ads turn as racist as a southern Democrat pining for the days when he could legally own people. And it'll probably happen about 5 minutes after the Googler decides it can run unattended. It'll probably be years before they figure out how to remove the racist rhetoric permanently, assuming it's even possible.
I think it'll take longer than people think before AI can start taking jobs.
It's important to approach the development and deployment of AI with responsible and ethical practices. While AI has shown great potential in various fields, including advertising, it's crucial to ensure that biases and inappropriate content are actively addressed and mitigated. Companies and developers are continuously working on improving AI systems to minimize the risks associated with biased or offensive outputs.
Regarding the impact of AI on jobs, it's a topic of ongoing discussion and research. AI has the potential to automate certain tasks and industries, leading to changes in the job market. However, it's also worth noting that AI can augment human capabilities and create new job opportunities. The exact timeline and extent of job displacement or creation are complex and can vary across industries.
As society progresses, it's essential to have ongoing dialogue and proactive measures to manage the impact of AI on the job market, including retraining and upskilling programs, ensuring equitable access to technology, and exploring new ways for humans and AI to collaborate effectively.
"public policy" is political, not an organic, realistic view of society.
Back in the 1950,s etc homosexuality was illegal for example.
If Google etc had existing back then, then these whole algorythms would have been "fudged" to reflect that, rather than the actual truth of the matter - as demonstrated by billions of unadulterated data points (users, commenst etc etc) .
Thats the risk / opportunity with AI - what its trained on...
Real world data, the truth as demonstrated by 98% of the population - or the "political agenda" of the day?
"Oooh Google! Ooooh, Big Data!"
So, as a consumer - i'm shown ads for things I may actually be interested in, based on things Ive actively shown an interest in in the past, and I'm taken to what is increasingly likely to be a good place for me to achieve what Im actually wanting to do?
My GOD! The horror!
Funnilty enough - and a lot of you pseudo-geeks on here - lets face it, most of your arent - might not get this bit...
Advertisers dont want to pay for advertising to people who arent interested in what tey sell
Customers dont wont to see ads for things their not interested in
Web servers and content etc that we all use and love (if not, why arr you looking at it?) need to be paid for
Websites and creators make money by providing a service or product that we as a customer value
I dont see an issue with any of that?
Now - GDPR and main stream media, where most of you on this rag of a site seem to get their "geek cool" credintionals from...
BBC and the press had 100% market share for decades - there was no other way of reaching a mass market
What you saw, what you didnt, what you know, what you dont...
Now, alomng comes Google and the internet...
Sh1t! People can now think for themselves, post what they want without a "Board" of 12 people - who of course have no agenda - deciding what that must be.
Market share - down
Control - down
Viewers - down
revenue - down
Now GDPR - a "law" created by a failing State purely to gernerate revenue from "Big Tech" as revenue from actual positive activities from the EU dries up - there never have been any.
To bolster that failing "control" weilded by the media, to control and stop you and I - the people - from actually engaging with each other, withouit the blinkers of "policy" and "agenda".
Now - why might the mainstream media, EU and "GDPR" have an isseue with that?
Jesus - wake up people.
..
We're likely not far away from the point where many consumers let algorithms do a lot of their discretionary shopping for them. Already we have plenty of popular "sell me stuff without making me choose" services, like those meal-preparation ones or Bespoke Post. I know people who have gotten random-clothing-of-the-month subscriptions and the like.
It's a small leap to just turning a budget over to a purchasing app that "learns your preferences". Or even not budgeting it, if it's reasonably successful at matching how much the victim consumer would have spent on unnecessary crap every time unit anyway. I wouldn't do it, but I'm a contrarian curmudgeon and I hate all joyful things such as spending money. You know there's no shortage of people who would. Every day is Christmas!
Then those machine-generated ads will gradually shift to targeting purchasing algorithms, rather than human purchasers, and we'll have an arms race between sell-bots and buy-bots. Which, again, should not be surprising; that's what we've seen happen in financial trading, for example, and in IT security. We can just take humans right out of the loop, and simply spend a lot of resources making and shipping crap around the world to be briefly admired and then tossed in a corner.
O brave new world (same as the old world).
And Why not?
Personally, after 2x Corona, I am extra fatiqued by choice and being presented for endless options and possibilites that I then have to specify Exactly to (maybe) get what I wanted.
For example: What I want is a nice barbecue with some friends, a few of which are vegetarian, with some drinks of varying yield and not ruinously expensive. For that, I would love to have a Star-trek interface, where the AI presents me with a long exposition and I can declare "Make it so".
Rich people employ valets and butlers to avoid all this stress and inconvenience, less rich people have PA's, and poor people can now have AI.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
