#AI#GenerativeAI#GeneratedImages#DevianArt#Bots#SocialMedia: "On March 27, a large group of artists and creators from across the web noticed the frightening extent to which a once-beloved, highly influential community platform of theirs had, like so many others, fallen prey to the artificial intelligence juggernauts plundering the internet.
As VFX animator Romain Revert (Minions, The Lorax) pointed out on X, the bots had come for his old home base of DeviantArt. Its social accounts were promoting “top sellers” on the platform, with usernames like “Isaris-AI” and “Mikonotai,” who reportedly made tens of thousands of dollars through bulk sales of autogenerated, dead-eyed 3D avatars. The sales weren’t exactly legit—an online artist known as WyerframeZ looked at those users’ followers and found pages of profiles with repeated names, overlapping biographies and account-creation dates, and zero creations of their own, making it apparent that various bots were involved in these “purchases.”
It’s not unlikely, as WyerframeZ surmised, that someone constructed a low-effort bot network that could hold up a self-perpetuating money-embezzlement scheme: Generate a bunch of free images and accounts, have them buy and boost one another in perpetuity, inflate metrics so that the “art” gets boosted by DeviantArt and reaches real humans, then watch the money pile up from DeviantArt revenue-sharing programs. Rinse, repeat.
After Revert declared this bot-on-bot fest to be “the downfall of DeviantArt,” myriad other artists and longtime users of the platform chimed in to share in the outrage that these artificial accounts were monopolizing DeviantArt’s promotional and revenue apparatuses. Several mentioned that they’d abandoned their DeviantArt accounts—all appearing to prove his dramatic point."
#AI#GenerativeAI#MidJourney#StabilityAI#GeneratedImages#WebScraping: "On Wednesday, Midjourney banned all employees from image synthesis rival Stability AI from its service indefinitely after it detected "botnet-like" activity suspected to be a Stability employee attempting to scrape prompt and image pairs in bulk. Midjourney advocate Nick St. Pierre tweeted about the announcement, which came via Midjourney's official Discord channel.
Prompts are the written instructions (like "a cat in a car holding a can of a beer") used by generative AI models such as Midjourney and Stability AI's Stable Diffusion 3 (SD3) to synthesize images. Having prompt and image pairs could potentially help the training or fine-tuning of a rival AI image generator model."
#AI#GenerativeAI#GeneratedImages#Copyright#AITraining#IP: "There’s a bit of a creative working ecosystem where there are natural cycles to someone getting their start to then having enough experience to gain a higher level in their career and make a living out of it. All of those doors are now starting to close because of generative AI. With the works that generative AI makes, the metric isn’t whether something is artistic or has quality, because if that were the case then people would win. The metric is the market itself. When you have a team trying to slash costs who say, “I could hire this artist full time, that’s a whole year’s worth of salary, or I could just pay a little subscription fee,” they’re going to go for the model.
Veterans who should be respected for the incredible contributions to our industry have been approached by high-profile production houses being like, “Can you paint over this Midjourney image? Oh, and we’ll pay you half.” That’s happening right now. At least in film, there’s at least some good pay a person can make a living off, and that’s now being lowered. And it’s going so much faster than any of us ever imagined. There’s a lot of angst and depression, even among actual professionals who are like, “I’ve given my whole life to this. It’s a lifetime of work.” And then for some company to say, “That lifetime of work, that dedication, it’s now mine. We’re gonna compete against you, we’re gonna make insane amounts of money off your work, and you don’t have to have a say.” That’s fucking a lot of people up. It’s a really tough time."
#AI#GenerativeAI#GeneratedImages#Instacart#Food: "An image of a recipe will not be this useful, I think, if it was AI-generated, and especially so if the fact that the image was AI-generated wasn’t disclosed by the recipe. That, to my surprise, is exactly the case with thousands of recipes the grocery delivery service Instacart is suggesting to its users. Some of the recipes include unheard of measurements and ingredients that don’t appear to exist.
Generally, I try to avoid using Instacart if possible because it treats its workers badly, but I had just come back from the hospital with a newborn and was desperate enough to pay a lot of money to have eggs and some other basics delivered to my doorstep. As I was browsing, I noticed that Instacart was offering me recipes that appeared to compliment the ingredients I was looking at.
The concept doesn’t make a ton of sense to me—I’m going to Instacart for the ingredients I know I need for the food I know I’m going to make, not for food inspo—but I had to click on a recipe for “Watermelon Popsicle with Chocolate Chips” because it looked weird in the thumbnail."
#AI#GenerativeAI#LLMs#AITraining#GeneratedImages#Copyright#IP: "A lot of early AI research was done in an academic setting; the law specifically mentions teaching, scholarship, and research as examples of fair use. As a result, the machine-learning community has traditionally taken a relaxed attitude toward copyright. Early training sets frequently included copyrighted material.
As academic researchers took jobs in the burgeoning commercial AI sector, many assumed they would continue to enjoy wide latitude to train on copyrighted material. Some feel blindsided by copyright holders’ demands for cash.
“We all learn for free,” Daniel Jeffries wrote in his tweet summing up the view of many in the AI community. “We learn from the world around us and so do machines.”
The argument seems to be that if it’s legal for a human being to learn from one copyrighted book, it must also be legal for a large language model to learn from a million copyrighted books—even if the training process requires making copies of the books.
As MP3.com and Texaco learned, this isn't always true. A use that’s fair at a small scale can be unfair when it’s scaled up and commercialized.
But AI advocates like Jeffries are right that sometimes it is true. There are cases where courts have held that bulk technological uses of copyrighted works are fair use. The most important example is almost certainly the Google Books case."
Who gains the most from this technology and through what mechanisms? Are benefits actually diffuse?
What new gatekeepers, intermediaries and consolidation of power is obscured?
Does equal access translate to equal or equitable outcomes among groups?
Whose priorities and preferences are embedded in the technology’s functionality?
There are no singular right answers here. Complex technologies defy singular interpretations of costs and benefits. Tradeoffs emerge between priorities like innovation, regulation, growth and equity. But too frequently, we fail even to probe beneath the superficial veneer of democratisation before adopting technologies at scale and dealing with unintended consequences after the fact."
#AI#OpenAI#GenerativeAI#TextToVideo#Sora#GeneratedImages: "Artificial intelligence company OpenAI showed off a new AI tool that can generate highly realistic 60-second videos based off a simple text prompt, a jump forward in quality for AI videos and “deepfakes” that have already been used to deceive voters.
The new tool, called Sora, will initially only be available to a small group of artists and filmmakers as well as “red teamers,” or researchers who try to find ways that an AI tool can be used for malicious purposes, OpenAI said in an announcement Thursday.
Sora builds on the tech behind OpenAI’s image-generating DALL-E tool. It interprets a user’s prompt, expanding it into a more detailed set of instructions, and then uses an AI model trained on video and images to create the new video."
This was the third time, in the past six months, that the iconic leader of the Dravida Munnetra Kazhagam (DMK) party was resurrected using artificial intelligence (AI) for such public events.
“When the COVID pandemic ravaged the world, our Chief Minister ran in the direction of panicked voices of people,” Karunanidhi said. “The nation knows the way you fought to save the lives of people, and so do I.”
Senthil Nayagam, founder of Muonium, the AI media tech firm that made the deepfake Karunanidhi video, told Al Jazeera that “there is a market opening up [for such deepfakes]…. You can attribute some statements to a particular person and that kind of gives more value to it”.
AI Karunanidhi’s first public appearance was at a local media event last year in September, which was followed up by another for a campaign by his party members. The resurrected leader often felicitates party workers and specifically praises the leadership of his son MK Stalin — with the aim of boosting his popularity."
#AI#GenerativeAI#GeneratedImages#Meta#Watermarking: "Meta is working to detect and label AI-generated images on Facebook, Instagram and Threads as the company pushes to call out “people and organisations that actively want to deceive people”.
Photorealistic images created using Meta’s AI imaging tool are already labelled as AI, but the company’s president of global affairs, Nick Clegg, announced in a blog post on Tuesday that the company would work to begin labelling AI-generated images developed on rival services.
Meta’s AI images already contain metadata and invisible watermarks that can tell other organisations that the image was developed by AI, and the company is developing tools to identify these types of markers when used by other companies, such as Google, OpenAI, Microsoft, Adobe, Midjourney and Shutterstock in their AI image generators, Clegg said." https://www.theguardian.com/technology/2024/feb/06/meta-ai-images-instagram-facebook-label-crackdown
#AI#GenerativeAI#GeneratedImages#China#Copyright#IP: "On November 27, 2023, the Beijing Internet Court (BIC) ruled in an infringement lawsuit (Li v. Liu) that an AI-generated image is copyrightable and that a person who prompted the AI-generated image is entitled to the right of authorship under Chinese Copyright Law (see our bilingual version, and the later-released official translation). Plaintiff generated an image of a woman by using Stable Diffusion, an open-source generative AI model that creates images from textual prompts. After publishing the disputed image on a Chinese social media platform (Xiaohongshu), Plaintiff discovered that Defendant had used the same image to illustrate an article on a different website without permission. Plaintiff then sued Defendant in the BIC.
Specifically, BIC made the following rulings:
The disputed AI-created image constitutes a “work” pursuant to the Copyright Law of the People’s Republic of China."
Sophie Maddocks: There’s a historian, Jessica Lake, and she’s done some really interesting research tracing the potential origins of the creation of fake nude images. She talks a lot about the rise of photography in the late 19th century, and writes about an example of face-swapping in late-19th-century photography where images of the faces of high-society women were pasted onto nude bodies and then circulated. And not only is that one possible starting point when thinking about the history of fake nudes, it’s also an interesting starting point for how we see the creation of A.I.–generated fake nudes. Fake nudes first went viral in the online sense in 2017 with the creation of the DeepNude app where the faces of individuals were digitally pasted onto the bodies of adult film actors, almost exactly mimicking what had been done in the late 19th century with photography.
So there is a long history to this harm, but I think there is that long-standing desire to produce fake nude images—almost exclusively of women. With the rise of the internet, we’ve seen ways of creating and sharing ever more photorealistic images—until we get to the last year with the rise of video- and image-generation models that create extremely realistic imagery and A.I. tools trained on millions of images of girls and women scraped from the internet without their consent. You can either use a text prompt or an existing image to produce a very realistic fake nude.
So A.I. has increased the volume and severity of this problem on the internet.
Absolutely. In 2017, when activists and the first people affected by A.I.–assisted deepfakes, like famous actors and singers, started to raise the alarm about this issue, they really gave us a roadmap for what would happen."
#AI#GenerativeAI#GeneratedImages#DeepFakes: "Microsoft has introduced more protections to Designer, an AI text-to-image generation tool that people were using to make nonconsensual sexual images of celebrities. Microsoft made the changes after 404 Media reported that the AI-generated nude images of Taylor Swift that went viral on Twitter last week came from 4chan and a Telegram channel where people were using Designer to make AI-generated images of celebrities.
"We are investigating these reports and are taking appropriate action to address them," a Microsoft spokesperson told us in an email on Friday. "Our Code of Conduct prohibits the use of our tools for the creation of adult or non-consensual intimate content, and any repeated attempts to produce content that goes against our policies may result in loss of access to the service. We have large teams working on the development of guardrails and other safety systems in line with our responsible AI principles, including content filtering, operational monitoring and abuse detection to mitigate misuse of the system and help create a safer environment for users.”"
#SocialMedia#Twitter#AI#GenerativeAI#GeneratedImages#Telegram: "Sexually explicit AI-generated images of Taylor Swift went viral on Twitter after jumping from a specific Telegram group dedicated to abusive images of women, 404 Media has found. At least one tool the group uses is a free Microsoft text-to-image AI generator.
Examples viewed by 404 Media had tens of thousands of bookmarks and likes and thousands of reposts. The Verge reported that one of the most viral examples received 45 million views and 24,000 reposts, and was up for 17 hours prior to its removal, when Twitter deleted the account of the original poster, as well as other accounts that posted similar images.
Swifties, being the highly motivated stan army that they are, mobilized to drown out the trending phrase with “protect Taylor Swift” tweets and buried many of the viral abuse images from the search results of the phrase "AI Taylor Swift" on Twitter. But other examples remain on Twitter as of writing.
404 Media found that the images originated in a Telegram group dedicated to making non-consensual AI generated sexual images of women."
#AI#GenerativeAI#GeneratedImages#Copyright#AITraining#FairUse: "The datasets on which GAI systems like ChatGPT and Stable Diffusion rely are more like Google Books than commercial web crawlers that gather data; GAI systems need writing and art in its complete form to train from. And the comparison to Google Books doesn’t stop there. Many GAI systems are built using third-party public datasets like Common Crawl and LAION that are distributed under fair use principles and provide a real public good: archiving and making accessible the aggregated content of the internet for academics, researchers, and anyone else that may want it. These are free, non-commercial datasets collected by nonprofit organizations for use by researchers and the public. Web crawling and scraping also underlie the operation of search engines and archiving projects like the Internet Archive’s popular Wayback Machine.
In other words, the same practices that go into collecting data for GAI training are currently understood to be non-infringing or protected by fair use. Considering how vital these practices are for an open and accessible internet, we should ensure that they stay that way.
As a threshold matter, it is critical to understand that accessing, linking to, or interacting with digital information does not infringe any copyright. Reading a book, looking at a photograph, admiring a painting, or listening to music is not, and never should be, copyright infringement. This is not a “fair use” issue; the ability to use, access, or interact with a creative work is outside a copyright owner’s scope of control. Based on the best explanations of how GAI systems work, training a GAI system is generally analogous to these kinds of uses."
#AI#GenerativeAI#GeneratedImages#Copyright#AITraining#IP: "Since the emergence of Midjourney and other image generators, artists have been watching and wondering whether AI is a great opportunity or an existential threat. Now, after a list of 16,000 names emerged of artists whose work Midjourney had allegedly used to train its AI – including Bridget Riley, Damien Hirst, Rachel Whiteread, Tracey Emin, David Hockney and Anish Kapoor – the art world has issued a call to arms against the technologists.
British artists have contacted US lawyers to discuss joining a class action against Midjourney and other AI firms, while others have told the Observer that they may bring their own legal action in the UK.
“What we need to do is come together,” said Tim Flach, president of the Association of Photographers and an internationally acclaimed photographer whose name is on the list.
“This public showing of this list of names is a great catalyst for artists to come together and challenge it. I personally would be up for doing that.”
The 24-page list of names forms Exhibit J in a class action brought by 10 American artists in California against Midjourney, Stability AI, Runway AI and DeviantArt. Matthew Butterick, one of the lawyers representing the artists, said: “We’ve had interest from artists around the world, including the UK.”
The tech firms have until 8 February to respond to the claim. Midjourney did not respond to requests for comment."
#AI#GenerativeAI#GeneratedImages#Datasets#AITraining: "Datasets are the building blocks of every AI generated image and text. Diffusion models break images in these datasets down into noise, learning how the images “diffuse.” From that information, the models can reassemble them. The models then abstract those formulas into categories using related captions, and that memory is applied to random noise, so as not to duplicate the actual content of training data, though it sometimes happens. An AI-generated image of a child is assembled from thousands of abstractions of these genuine photographs of children. In the case of Stable Diffusion and Midjourney, these images come from the LAION-5B dataset, a collection of captions and links to 2.3 billion images. If there are hundreds of images of a single child in that archive of URLs, that child could influence the outcomes of these models.
The presence of child pornography in this training data is obviously disturbing. An additional point of serious concern is the likelihood that images of children who experienced traumatic abuse are influencing the appearance of children in the resulting model’s synthetic images, even when those generated images are not remotely sexual.
The presence of this material in AI training data points to an ongoing negligence of the AI data pipeline. This crisis is partly the result of who policymakers talk with and allow to define AI: too often, it is industry experts who have a vested interest in deterring attention from the role of training data, and the facts of what lies within it. As with Omelas, we each face a decision of what to do now that we know these facts."
#AI#GenerativeAI#GeneratedImages#Copyright#IP: "CS: As artists: what do we want from AI? Are there concrete actual requirements for the AI community? Consent, credit and compensation I think is a major thing for me. I wish we could have nuanced conversations about Generative AI without it sounding alarmist, though I’m cognizant of the fact that I do think that this is going to impact aspects of the creative industry. I wouldn't be surprised if we see smaller and smaller fellowships, or a lot more confusion over the artistic and creative practice. Underpinning the global conversation are the misconceptions of how much creativity is worth and how long it takes to be creative, how long it takes to work on a piece of art. This capitalism hellhole we're in: every hour is subdivided into billable minutes, and with the rise of the gig economy, suddenly there’s an expectation that you’ll create a piece of really good work, that you are underpaid for, really quickly. You need a lot of time to be creative. You just need time to sit down and stare at a screen sometimes or look at something that has nothing to do with your work. That is part of the process. And I worry that this further flattens that, and puts us in a place where we're all urgent all the time."
#UK#AI#GenerativeAI#GeneratedImages#StabilityAI#GettyImages#Copyright: "The core of the claimants’ allegations is that Stability AI scraped millions of images from the Getty website without consent, and used these images to train and develop Stable Diffusion. It’s further claimed that the synthetic images generated by Stable Diffusion, accessed by users in the UK, infringe upon Getty Images’ copyrighted works and bear their trade marks. Some of these images had been presented in the particulars of claim, but it was never made clear how the
images came to be. I was able to produce some images myself with older versions of Stable Diffusion bearing the semblance of a Getty Images logo, but none of the outputs produced appeared to come from the images in the input. The idea here is that Stable Diffusion “memorised” the Getty logo, and could place it on outputs on demand. This is no longer possible as far as I can tell."
#AI#GenerativeAI#GeneratedImages#Emu: "Meta's model generally creates photorealistic images well, but not as well as Midjourney. It can handle complex prompts better than Stable Diffusion XL, but perhaps not as well as DALL-E 3. It doesn't seem to do text rendering well at all, and it handles different media outputs like watercolors, embroidery, and pen-and-ink with mixed results. Its images of people seem to include diversity in ethnic backgrounds. Overall, it seems about average these days in terms of AI image synthesis."
#AI#GenerativeAI#GeneratedImages#Copyright#IP: "In the amended complaint filed this week, the original defendants are joined by seven additional artists: Hawke Southworth, Grzegorz Rutkowski, Gregory Manchess, Gerald Brom, Jingna Zhang, Julia Kaye, Adam Ellis.
Rutkowski’s name may be familiar to some readers of VentureBeat and our colleagues at GamesBeat as he is an artist from Poland known for creating works for video games, roleplaying games, and card games including the titles Horizon Forbidden West, Dungeons & Dragons, and Magic: The Gathering.
As early as a year ago, Rutkowski was covered by news outlets for complaining that AI art apps based on the Stable Diffusion generation model were replicating his fantastical and epic style, sometimes by name, allowing users to generate new works resembling his for which he received zero compensation. He was also not asked ahead of time by these apps for permission to use his name.
Yesterday, Rutkwoski posted on his Instagram and X (formerly Twitter) accounts about the amended complaint, stating “It’s a freaking pleasure to be on one side with such great artists.”"
#AI#GenerativeAI#GeneratedImages#Energy: "Each time you use AI to generate an image, write an email, or ask a chatbot a question, it comes at a cost to the planet.
In fact, generating an image using a powerful AI model takes as much energy as fully charging your smartphone, according to a new study by researchers at the AI startup Hugging Face and Carnegie Mellon University. However, they found that using an AI model to generate text is significantly less energy-intensive. Creating text 1,000 times only uses as much energy as 16% of a full smartphone charge.
Their work, which is yet to be peer reviewed, shows that while training massive AI models is incredibly energy intensive, it’s only one part of the puzzle. Most of their carbon footprint comes from their actual use."
#AI#GenerativeAI#GeneratedImages#Journalism#Media#News: "The AI authors' writing often sounds like it was written by an alien; one Ortiz article, for instance, warns that volleyball "can be a little tricky to get into, especially without an actual ball to practice with."
According to a second person involved in the creation of the Sports Illustrated content who also asked to be kept anonymous, that's because it's not just the authors' headshots that are AI-generated. At least some of the articles themselves, they said, were churned out using AI as well.
"The content is absolutely AI-generated," the second source said, "no matter how much they say that it's not."
After we reached out with questions to the magazine's publisher, The Arena Group, all the AI-generated authors disappeared from Sports Illustrated's site without explanation."
#UK#AI#GenerativeAI#GeneratedImages: "Children in British schools are using artificial intelligence (AI) to make indecent images of other children, a group of experts on child abuse and technology has warned.
They said that a number of schools were reporting for the first time that pupils were using AI-generating technology to create images of children that legally constituted child sexual abuse material.
Emma Hardy, UK Safer Internet Centre (UKSIC) director, said the pictures were “terrifyingly” realistic.
“The quality of the images that we’re seeing is comparable to professional photos taken annually of children in schools up and down the country,” said Hardy, who is also the Internet Watch Foundation communications director.
“The photo-realistic nature of AI-generated imagery of children means sometimes the children we see are recognisable as victims of previous sexual abuse."
A group of researchers managed to get both Stability AI’s Stable Diffusion and OpenAI’s DALL-E 2 text-to-image models to disregard their policies and create images of naked people, dismembered bodies, and other violent and sexual scenarios.
Their work, which they will present at the IEEE Symposium on Security and Privacy in May next year, shines a light on how easy it is to force generative AI models into disregarding their own guardrails and policies, known as “jailbreaking.” It also demonstrates how difficult it is to prevent these models from generating such content, as it’s included in the vast troves of data they’ve been trained on, says Zico Kolter, an associate professor at Carnegie Mellon University. He demonstrated a similar form of jailbreaking on ChatGPT earlier this year but was not involved in this research."
#a16z#VCs#AI#GenerativeAI#GeneratedImages#Deepfakes: "Andreessen Horowitz, also known as a16z, the influential Silicon Valley venture capital firm that was an early investor in Facebook, Lyft, and other tech giants, has invested in Civitai, a giant platform for sharing AI models that enables and profits from the creation of AI generated nonconsensual sexual images of real people. That includes launching a feature where people can list “bounties” for others to create AI models of specific targets.
Civitai said that it raised $5.1 million in a seed funding round led by a16z.
A16z’s official website, which includes a jobs board with open positions at companies in its portfolio, currently lists five jobs at Civitai. According to a16z’s site, these jobs were posted more than 30 days ago."