Discover more from GOOD INTERNET
The ChatGPT defamation lawsuits cometh
Plus: Elon Musk lacks critical bodypart / Clickbait Programming Language / How To Blow Up A Pipeline is a movie / ChatGPT identifies Shrek / Indy 5 / Casual Everyday Deathmetal / and much more
When stochastic Gossip-AI hallucinates false accusations about real people
In one of my first posts after the launch of AI-image generator Stable Diffusion was about the legal pitfalls of image synthesis. This was some months before the lawsuits against Microsofts Copilot and Stable Diffusion and, in the face of these copyright-lawsuits, i compared the current wave of generative AI to the Napster-age of filesharing.
Now chatbots are in legal hot water, too: A lawyer asked ChatGPT to produce a list of colleagues who sexually harassed someone. ChatGPT, in good old stochastic parrot fashion, was happy to oblige and produced a list that wrongly featured the name of law professor Jonathan Turley. Here’s a WaPo-piece on that incidence: “The AI chatbot fabricated a sexual harassment scandal involving a law professor — and cited a fake Washington Post article as evidence”. It is important to understand that ChatGPT was not explicitly prompted to write a paragraph about the crime history of that law professor, but simply hallucinated his name in a list of criminals.
In a similar case, an australian mayor is readying a defamation lawsuit against OpenAI because it made up facts about a criminal conviction that never happened.
Now, if i'd spread lies about anyone, a law professor is the last one i'd target. Besides copyright issues, defamation-lawsuits from chatbots hallucinating false Gossip about real people is a real danger to AI-development, and maybe it's a good thing it hit a law professor first. He surely knows how to handle it, and if i were the Washington Post, i'd too talk to some lawyers now.
I'm not sure how AI-companies could prevent robots hallucinating false gossip about real people, ChatGPT and Bing picking up wrong gossip from the web or simply making up legally relevant facts, especially with internet access. In a new study, researcher found that co-writing with opinionated language models affects users' views. This surely holds true when ChatGPT inclused your name on a made up list of sexual predators and it's one thing when people talk shit about you, it's another when companies help spreading it.
In germany, we have shady lawyers we call “Abmahnanwalt“, a guy who uses legal loopholes to produce cease and decist-letters and accompanying invoices en masse. This phenomenon was very relevant for filesharing a decade ago, and with AI-text-synthesis vultures like these can flood any AI-company with a mass of complaints. But that’s just one side of this story.
Many will point at “AI is a tool“ and that you get what you prompt, so when you prompt for an article about “Person Y sexually harassed person X“, it will do so and what do you expect? But this is not the same as a person typing a defamatory text in a word document, it is not a human who produces that defamation, it’s a server in San Francisco owned by OpenAI.
And, while this issue is only tangential related to the privacy complaints leading to Italy banning ChatGPT, with France, Germany and Ireland considering similar moves and Canadas Privacy Commissioner opening an investigation, i do think that AI hallucinating wrong histories of people is an invasion of privacy on an unpredecented level.
Given that these stochastic-gossip-models are integrated into the biggest text-interfaces in the world including search and Office-products, Microsoft and OpenAI are in trouble and, you know, in a race, when you drive too fast, physical forces will drag your car into the ditch. This also may hold true for AI arms races and Microsoft and OpenAI gonna find out that for AI companies, too, the real world moves slower than you think.
Elon Musk has no balls
In an unexpected turn of events, Elon Musk confirmed that he has, indeed, no testicles, no balls, a complete lack of reproductive organs. A press release from Twitter HQ reads: “It is true, our CEO has no balls. Never had.“
Doge Coin went down in flames to the news that, wow, yes, so concerning, Elon Musk has no family jewels at all.
The news came to light when Tesla Employees shared private images of their CEOs lower area shot with an onboard cam while driving a robot car.
While speaking at the international congress for hair transplants, he adressed the issue: “Life without testicles is hard, folks, i hope you understand.“ We do, and feel sorry for the richest man in the world who not only owns the hivemind platform, but also the robot car platform, and the brainreading platform, and the rocket platforms for reusable space exploration things, and who has no nuts.
Testicles are enclosed in the scrotum, and it is unclear at this point of reporting if Elon Musk is lacking the whole whiggling sack of private parts, or just the innards. Whatever the case: Elon Musk has no balls.
Spokespersons from Mastodon and Substack in a joint statement commented on this lack of organ: “We think that Elon Musk should grow a pair“. Elon Musk personally blocked Links to both platforms on Twitter in retaliation.
Matt Taibi, who famously reported extensively on the Twitter Files before, commented on the missing testicles by confirming that he, indeed, does own a pair of big hairy balls, which was immediately validated by the fact of Musk, who has no balls, unfollowing his former buddy on the Tweeties.
The author of this small piece of satire will take a good look at the ballsy Notes-feature coming to Substack soon and start to use it if cool. Needless to say, he will continue to not take every childish tantrum of powerful platform-owners, or himself, too seriously.
Tabloid - A Clickbait Programming Language
Shocking New Programming Language Bewilders Programmers at Google and Facebook! The Secret Programming Language Every 10x Programmer Recommends! Your Friends Will Be Jealous About This New Programming Language!
Tabloid is a turing-complete programming language for writing programs in the style of clickbait news headlines.
a few thingsthe Top Five Most Popular Quirks and Features of the Tabloid programming language (Number Four Will Shock You!)
Print output with the keywords
YOU WON'T WANT TO MISSfollowed by an expression. Everything printed by Tabloid is automatically capitalized, and an exclamation point is added. Why would you want anything else?
Declare a function by writing
DISCOVER HOW TO ... WITH. Truly, a more gripping way to declare a function can't possibly exist! Similarly, assign to a variable with
EXPERTS CLAIM ... TO BE. On the Internet, anyone can be an expert, and Tabloid gives YOU the power to wield that responsibility and declare anything you'd like!
There are no built-in constructs for looping. The news cycle is moving too fast! Nobody has time for yesterday's loops or last week's break statements. If you must loop, use recursion.
This is the best eso programming language since Rockstar for hair metal coding.
How To Blow Up A Pipeline is a movie, and conservatives will love it, i’m super very sure about that.
Last year in germany, conservatives threw a tantrum when climate activist Luisa Neubauer quoted the title of a book called How To Blow Up A Pipeline at a press conference. Some months later, conservatives still referenced this non-issue as a marker of radicalization of the climate movement when activists threw soup on glas-shields protecting art, or glued their hands to the streets, blocking them. Many conservative commentators used these instances to point at the alleged illegality of more radical climate protests.
That book is a movie now, here’s the trailer, and I want to use that opportunity to get something out of my system.
In a historic legal decision, “the Dutch Supreme Court, the highest court in the Netherlands, upheld the previous decisions in the Urgenda Climate Case, finding that the Dutch government has obligations to urgently and significantly reduce emissions in line with its human rights obligations.”
In April 2021, the german supreme court, in a similar decision, pointed at a principle of generational justice: One generation should not be allowed to suck up all CO2-Budget and shift a radical burden of carbon-reduction on following generations, simply because that’s an unjust limitation of freedom guaranteed by the law. For the german supreme court, not fighting climate change is illegal. Similar lawsuits are ongoing in states like the US, Belgium, Sweden, India, New Sealand, Norway, South Africe, the Swiss and, I hope, many more to come.
I view radical climate protest in the context of these lawsuits, with trangressions during these protests to be decided on a case by case basis. But the argument, that these protests which are blocking streets, try to stop the ecivtion of a whole village for a new coal mine, or disturb the delicate enjoyment of art (while making sure that the artworks themselves are not damaged), don’t stand on legal ground per se, is nil in the face of these decisions by the supreme courts. Any good lawyer defending more radical climate activism in court will point at those decisions and drive home the point that these kids are actually upholding and defending the law.
And in that context, it baffles me that conservatives are ignoring legal decisions, but consistence has never been a property of conservative thinking, i guess.
I’m not sure if the pipeline movie will be a new Die fetten Jahre sind vorbei (eng. The Edukators), in which radical anarchists kidnap a businessman, who then goes on to explain some philosophical inconsistencies in their ideological thinking, form a bond with the guy to finally work together against an unjust system. But such a thing is dearly needed, even when i am, myself, far from being a radical or even a true anticapitalist.
Either way, I’ll surely watch and enjoy the hell out of this movie.
Who Owns SpongeBob? AI Shakes Hollywood’s Creative Foundation: Hollywood is scratching their heads giving the interpolative character of generative AI: Who owns Batman in a Pokemon-suit?
ChatGPT is making up fake Guardian articles. Here’s how we’re responding — There’s not much about their response there, but i’m very sure their legal department is involved.
Related: Metaphysic CEO files for copyright of his own AI likeness — Copyright for people in the age of mimetic AI. If this goes through, this should set a new standard for all kinds of cloning AI-tech (voice, deepfakes etc). Voice actors already are being asked to sign away the rights to their voice for AI-training. A copyright for human likeness may be a solution for this. Also: Character.ai, the company producing mimetic AI-models of famous people, may have a problem with this.
Also related, here’s a remarkable piece of Voice-Cloning AI: Hey There Delilah but it's Kanye's Voice.
Artificial intelligence in communication impacts language and social relationships: “We find that using algorithmic responses changes language and social relationships. More specifically, it increases communication speed, use of positive emotional language, and conversation partners evaluate each other as closer and more cooperative. However, consistent with common assumptions about the adverse effects of AI, people are evaluated more negatively if they are suspected to be using algorithmic responses. Thus, even though AI can increase the speed of communication and improve interpersonal perceptions, the prevailing anti-social connotations of AI undermine these potential benefits if used overtly.“ Writeup on Neuroscience News.
AI-Researchers at the University of York have found that current AI-generated music is inferior to human-composed music. We’ll see how long this holds up.
Open Letter from the dutch AI Summer School: We are not ready for manipulative AI
How Will AI Transform Photography? — Good piece on post-photography in the age of AI: “Our physical gestures are expressive of internal, psychological states, but AI struggles to process the aesthetic of emotions. Grief or pleasure may appear on AI-generated faces but isn’t replicated in those figures’ postures or gestures. Engman has observed that the body language of performers includes subtle movement choices cultivated over time to express thoughts and feelings, but these are rarely read accurately across AI data sets. Tags associated with images don’t typically specify a relationship between affect and a particular gesture. For instance, an emotion might be determined as happy because many images with smiles are tagged ‘happy’, but the AI might not be prompted to discern other subtle postures or stances, such as relaxed shoulders. For Engman, this gap is a compelling reason to explore the technology.“
GPT-4 tries to design a universal Turing Machine using a standard Ouija board giving the term “Ghost in the machine“ a whole new meaning.
Life after Daft Punk: Thomas Bangalter on ballet, AI and ditching the helmet: “The last thing I would want to be, in the world we live in, in 2023, is a robot.“
The Case for Banning Children from Social Media — Social Media are drugs and should be regulated in a similar way.
Galaxy Simulation: “This project is a simulation of thousands of particles that reproduce gravity to simulate galaxies. It uses ThreeJS for 3D rendering and shaders to calculate the movement of particles in real-time.“
I’m not big on board games, but i did play my fair share of Settlers of Catan some years ago. So, RIP Klaus Teuber, Creator of the Board Game Catan, and thanks for all the fun.
Ten iconic directors shoot the World Cup — Football sucks, but this is great.
Mony Lisa never looked better, from Shardcore on Twitter:
The Dawn of Twitter by Roland Meyer:
I miss the superfreak days of AIart, but Midjourney surely is a class of its own:
GOOD INTERNET is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.