On Wednesday, web infrastructure provider Cloudflare announced a new feature called "AI Labyrinth" that aims to combat unauthorized AI data scraping by serving fake AI-generated content to bots. The tool will attempt to thwart AI companies that crawl websites without permission to collect training data for large language models that power AI assistants like ChatGPT.
Cloudflare, founded in 2009, is probably best known as a company that provides infrastructure and security services for websites, particularly protection against distributed denial-of-service (DDoS) attacks and other malicious traffic.
Instead of simply blocking bots, Cloudflare's new system lures them into a "maze" of realistic-looking but irrelevant pages, wasting the crawler's computing resources. The approach is a notable shift from the standard block-and-defend strategy used by most website protection services. Cloudflare says blocking bots sometimes backfires because it alerts the crawler's operators that they've been detected.
(Score: 5, Funny) by Barenflimski on Sunday March 23, @07:16AM (4 children)
This is the same thing that Reddit does -> "turns AI against itself with endless maze of irrelevant facts"
(Score: 3, Interesting) by Anonymous Coward on Sunday March 23, @11:52AM (3 children)
"turns AI against itself with endless maze of irrelevant facts"
Doesn't "Outbrain", "Taboola", and similar click bait sites do this exact same thing already to people? They tantalize one with something and tease but won't offer the tidbit, teasing me if I just click yet another link that they will reveal, yet I have just spent five minutes going through pages of irrelevant verbiage with no reward. I know by this time I am not dealing with anything useful...just a click bait trap over why I should place a toilet paper roll under the seat. I have wised up now and just DDG it and usually get a concise answer, usually making no sense, but at least I didn't have to click through 100 pages to see I've been had for a gullible fool that clicks on this sort of thing.
Everytime I see their name show up, from experience, I recognize the same MO and abort.
(Score: 3, Insightful) by mcgrew on Sunday March 23, @03:47PM (2 children)
Sucker!
How many of them caught you before you wised up? Snopes speaks of them a lot. Now thanks to the huge number of gullible people like you, the God damned newspapers are doing the same fucking thing; a Google News headline sends you to Newsweek or Time or a hundred other mainstream sites that don't show what is advertised, but want you to subscribe!
The dumbass outlets pulling this nonsense won't get my money, I refuse to do business with the dishonest. Why would I pay to get information from liars? Of course, there's so much dishonesty in this dystopian century that it's hard to find an honest business to shop in, or honest news outlet to listen to. I miss the days before "infotainment". An actor shouldn't be on the news unless he's running for governor or she invents spectrum hopping.
Impeach Donald Saruman and his sidekick Elon Sauron
(Score: 1, Insightful) by Anonymous Coward on Monday March 24, @12:13AM (1 child)
What they did is give me an irreversible distrust of all media. Google News Headlines are trusted about the same as telemarketing calls. People like Mike Adams ( Brighteon.com ) or Alex Jones ( InfoWars ) come across with more credibility than Google News or the mainstream media. The Masters Degree in Marketing people monetized the reputation of the corporate that hired them, irreversibly soiling the company reputation by overemphasizing short term goals. It became Wham! Bam! Thank you, Ma'am!". Eats, shoots, and leaves.
Once the Marketing Guys are hired, corporate reputations soon become monetized into executive bonuses, and the spent shell of the corporation that hired them is tossed with all the care of a tossed beverage cup once it's content has been drained.
I have seen America's Top Manufacturing concerns bite the dust business concerns override the integrity of the product, however these same business paradigms work well in finance, marketing, rent-seeking, government, services, well anything that doesn't make anything.
We have become a nation of useless eaters who convert natural resources into useless displays of wealth. It puzzles me why this system I am in continues to function at all. Seems to me we are living "high on the hog" off an inheritance, which when we exhaust it, things will get desperate ( and for many, that is already the case ).
(Score: 1) by khallow on Tuesday March 25, @12:01PM
If they do come across with more credibility, then it's probably some sort of disease. The brighteon.com thing doesn't even appear to be news. At a glance, I can't even tell what they're selling - the site is a mess. As to Alex Jones, he got caught and successfully sued for defaming Sandy Hook victims and their families for years. These aren't the media sources that will be taking on Google News Headlines.
(Score: 2) by HeadlineEditor on Sunday March 23, @09:57AM (11 children)
This is and will be the best use of AI technology ever.
(Score: 0) by Anonymous Coward on Sunday March 23, @10:02AM (1 child)
Do you really need AI for this though? Something that uses lower CPU etc might be able to poison AI as well.
(Score: 4, Touché) by HeadlineEditor on Sunday March 23, @11:12AM
You're absolutely right - this is a terrible waste of energy. A better use would be to electrocute the developers of AI "assistants." But for some reason we're not allowed to do that.
(Score: 2) by mcgrew on Sunday March 23, @03:55PM (7 children)
I see you know very little about AI. It's more than chatbots these days, real scientists in many fields are using it for real science and getting excellent results.
AI now is used in microsurgery; computers don't shiver or sneeze. My mom had surgery done by a robot that would have been impossible for a surgeon with a scalpel. She died ten years later at age ninety two.
That said, computers still don't think and can't do anything creative without human brains doing the true creativity.
Impeach Donald Saruman and his sidekick Elon Sauron
(Score: 3, Insightful) by Anonymous Coward on Sunday March 23, @05:55PM (3 children)
then stop calling it "AI"
(Score: 0) by Anonymous Coward on Monday March 24, @02:19AM
> then stop calling it "AI"
I call it advanced pattern matching, or perhaps a fancy neural net. Also, I'm careful to not anthropomorphize this technology, staying away from words that describe human behavior like "think" and "hallucinate".
In like manner I never say something is "in the cloud", since it actually is "on someone else's computer."
While easy to fall for the hype, you have a choice to do your own thinking/research, and see things as they are.
(Score: 2) by janrinok on Monday March 24, @07:49AM (1 child)
This is very much the same argument as the use of the word "hacker". We are much too late and the MBAs have managed to convince the majority of people that the AI name is the most appropriate.
We can sit and complain over our coffee that it is incorrect, and that it is a misleading and inaccurate term. The only thing that will change will be the temperature of the coffee...
I am not interested in knowing who people are or where they live. My interest starts and stops at our servers.
(Score: 0) by Anonymous Coward on Monday March 24, @12:24PM
Maybe we should conflate them. Every chance you get, refer to them as AI hackers.
(Score: 5, Insightful) by https on Sunday March 23, @06:28PM
Thanks to incessant hype from the bastards who have to sell something before the venture capital dries up, AI now means LLM chatbot. Expert systems/deductive solvers, automatic classification systems, annealing, ant/genetic algos... all swept away.
Kind of like how now, when you say "hacker" everybody except programmers over a certain age thinks "criminal" - though to be fair that wasn't VC, just morons.
Offended and laughing about it.
(Score: 2) by ledow on Monday March 24, @10:09AM (1 child)
Robotics and automation - have been around for nearly a century, if not longer.
AI doesn't exist.
What we have are advanced spam filters running around pretending to be human and failing even more miserably than humans.
Nobody has a problem with robot surgery, or an "AI" - really just an image analyser - scanning your X-ray for cancerous lumps or whatever.
But it's NOT AI.
(Score: 2) by mcgrew on Monday March 24, @03:53PM
Robotics and automation - have been around for nearly a century
The first robot was Unimate from the early 1950s; the first computer, ENIAC, was patented in 1946. Unamate's inventor couldn't sell it until the 1960s. Automation is about sixty years old.
AI doesn't exist.
AI is just a meaningless name for a class of things giant computers do. In 1952's presidential election, CBS News introduced computers to the world. [npr.org] It was called an "electronic brain" and was less powerful than a Hallmark greeting card.
Impeach Donald Saruman and his sidekick Elon Sauron
(Score: 2) by corey on Sunday March 23, @10:14PM
Agree. But I can’t help but feel sad that on the whole, we’re burning energy away to burn energy away. It’s like having two electric motors driving against eachother.
(Score: 3, Interesting) by anubi on Sunday March 23, @10:38AM (3 children)
Sounds like this is the Cloudflare response to bots as Lenny is for telemarketers. I often put Lenny's phone number in when someone puts a required phone number on their web form.
Lenny's Public Number : ( 423 ) 468 4041
( Lenny is a bot. His sole mission is entertain telemarketers ). I just checked the number to make sure it was Lenny's, as Lenny welcomes calls at all hours.
https://m.youtube.com/playlist?list=PLduL71_GKzHHk4hLga0nOGWrXlhl-i_3g [youtube.com]
I thought maybe some of us might get a kick out of this... Such a simple mouthtrap, but sure is hilarious when he catches one.
"Prove all things; hold fast that which is good." [KJV: I Thessalonians 5:21]
(Score: 2) by mcgrew on Sunday March 23, @04:06PM (2 children)
Do you have one in the 217 area code? Everybody needs a Lennie and it would have to be an Illinois number at the weed dispensary. It should be illegal to require a phone number without true need.
I've been using (217) 3825-633. I hope that's not a real number, it's (217) FUCK-OFF.
Back in the 33.6 days I bought a modem with some sophisticated answering machine software (it was also the landline days). I made a labyrinth for telemarketers with it, a number not listed friends and family knew let it ring. Today I just don't answer unknown numbers, there's voicemail and text if your phone isn't in my phone's directory.
Impeach Donald Saruman and his sidekick Elon Sauron
(Score: 1) by anubi on Monday March 24, @11:06AM (1 child)
Sorry, that's the only Lenny I have.
Here's his home page. Sure is a likeable old guy...until you try to get him to commit to anything.
https://lenny.penguintek.net/ [penguintek.net]
"Prove all things; hold fast that which is good." [KJV: I Thessalonians 5:21]
(Score: 1) by anubi on Monday March 24, @10:19PM
Just found another British Telecom AI that does one better than Lenny...
https://m.youtube.com/watch?v=RV_SdCfZ-0s&t=101s&pp=2AFlkAIB [youtube.com]
It's a nice elderly grandma, with AI video and even cats!
"Prove all things; hold fast that which is good." [KJV: I Thessalonians 5:21]
(Score: 1, Troll) by SomeGuy on Sunday March 23, @01:32PM
So this is why everything on Soylentnews has looked like watered down gibberish lately :P
But seriously, given cloudsnare's recent behavior blocking non-mainstream browsers, everyone should fully expect to be mistaken for a bot eventually.
It wouldn't surprise me if I have already encountered this but just assumed it was yet another clickbait or paywall site.
(Score: 2) by https on Monday March 24, @02:01AM
Instead of feeding them slop, which requires running a slop generator locally or downloading slop from elsewhere (which sends the unfortunate message "we like slop" to the perpetrators), feed them bad information guaranteed to get them in shit:
E.g. pictures of cumshots, named like "cat_17.jpg" and with the text "meow" as overlaid caption. You know, random stills from the last five minutes of whatever it is that has cumshots in the last five minutes. Those shouldn't be too hard to find, and easily processed with bash, perl, and ffmpeg.
It turns out, the existance of particular books (and maybe other media) in training data is verifiable. So I am wondering about the legality of having a copy of say, Harry Potter or Metallica's "...and Justice for All", or the CRC Materials Handbook or whatnot, hidden behind a robots.txt - it could be argued, "no, that is not distributed - that is for my PERSONAL use, and there are no public links to it.
Offended and laughing about it.