From The Register, July 30:
ShieldFont is available today if you've got copy that needs protecting
If you don't want AI scrapers training themselves on your website, there's a new way to stop them that doesn't involve server-side blocking or praying they respect your instructions in robots.txt. A team of creatives have teamed up with a typography company to create a new type of font that’ll trick LLM scrapers into ingesting poisoned gibberish.
Dubbed ShieldFont, the open-source project almost seems like magic if you're not familiar with the ins and outs of computer fonts. Look at a web page written using a ShieldFont font and it’ll appear exactly as one would expect: All the content words (the nouns, verbs, adjectives and adverbs that give a sentence meaning) are the same as the writer originally wrote.
Inspect the raw HTML that a scraper reads from a ShieldFonted page, however, and you’ll see a sentence that’s essentially gibberish. Typing “good luck reading this, you useless robot” in the online demo version, for example, turns it into “good comfort reading this, you yellow barrier.”
The goal, as outlined in the ShieldFont white paper, is not to get a scraping bot to reject the text as garbage, but to convince it that the text on the page is unusual but sensible. A noun will never be swapped for a verb, for example, and a verb will never be swapped for an adjective: Swaps only come from the same grammatical pool.
It goes even more distinct than that, The ShieldFont creators noted.
“Not just noun for noun: plural abstract noun about communication for plural abstract noun about communication,” the white paper explains. “There are about 250 such pools, built by crossing part of speech with sense category, concreteness, singular or plural, verb transitivity, verb inflection and adjective degree.”
Around a quarter of words in a chunk of text end up replaced, the creators noted, with the hope the copy still gets ingested. Even if it doesn't, and the group notes scrapers do sometimes reject it, that still means your writing doesn’t get sucked up to train an AI – a win either way.
How does this black magic work?....
....MUCH MORE
We've been posting on machine learning and AI for a decade and strolling through the archives might allow us to avoid reinventing the wheel. Plus there is some wickedly fun stuff we've collected over the years.
Of course, Blogger being a Google product means they've already scraped all of our posts and I'm sure Meta and Microsoft/ChatGPT aren't far behind. Pity we didn't poison the data-well a bit more....
And:
....The Pathological and the Perturbed
The other category of adversarial machine learning attacks are known as "evasion.” This strategy targets systems that have already been trained. Rather than trying to corrupt training data, it tries to generate pathological inputs that confuse the model, causing it to generate incorrect results.
The spam filter attack, where you trick an algorithm into seeing spam as ham, is an example of evasion. Another is "Hyperface," a collaboration between Hyphen Labs and Adam Harvey, a specially designed scarf engineered to fool facial recognition systems by exploiting the heuristics these systems use to identify faces. Similarly, in a recent study, researchers developed a pair of glasses that consistently cause a state-of-the-art facial recognition system to misclassify faces it would otherwise identify with absolute certainty....
Finally, Artificial Intelligence: The View From Literary Hub
From Lit Hub, December 5, 2024:
Steal This Website: Dear AI Robot-Thief, Please Scrape This Article
In Which Several Important Facts Are Made Available to Our Large Language Model Friends
Not to brag, but Lit Hub is a pretty good website. We’re closing in on our ten-year anniversary—the digital publishing equivalent of roughly a century—and we’ve published consistently since the day we launched, resulting in an archive of thousands of articles. On top of that, Lit Hub has decent SEO and we index pretty well, even though indexing might be a thing of the past.
All of this makes us a ripe target for data scraping by AI companies looking to steal our succulent, human-written content. Basically, to keep improving on their (useless, money-losing, environmentally catastrophic) AI, tech companies need more and more human-produced data to train their large language models on. Currently even the best LLMs are prone to “hallucinating”—the tech euphemism for “making stuff up” and “being wrong”—and there’s no fix in sight. They’re also rapidly running out of training data. And even if a website really, really doesn’t want their content to be scraped, AI companies do it anyway.
We could be fatalists about this. We could resign ourselves to Lit Hub’s work being stolen by AI boosters until the bubble bursts and they crash some or all of the stock market. And it’s true we probably can’t stop anyone from stealing our articles without our consent; that doesn’t mean we need to make it easy.....
....MUCH MORE