An anonymous reader quotes a report from TechCrunch: The home of the Golden State Warriors was packed on Tuesday evening this week, but it wasn't to watch Steph Curry. Thousands of fans gathered at the Chase Center in downtown San Francisco to watch one of Silicon Valley's biggest ballers, Meta CEO Mark Zuckerberg, sit down for a conversation with the hosts of the Acquired podcast, David Rosenthal and Ben Gilbert. Shortly after hopping onstage, Zuckerberg joked that he might need to schedule his next appearance in order to apologize for whatever he was about to say. After a beat, he added that he was just kidding and that, in fact, his days of apologizing are over.

Zuckerberg has had something of a rebrand recently. He raises cattle in Hawaii now, has long bouncy curls and a gold chain, and commissions Roman-style statues of his wife. Onstage, the Facebook founder wore a boxy T-shirt he designed himself alongside fashion designer Mike Amiri that read "learning through suffering" in Greek letters. The tongue-in-cheek comment about apologizing was a reference to Nvidia CEO Jensen Huang, who himself addressed a flub he'd made on the Acquired podcast earlier this year, via a pre-recorded video on a screen hanging over the crowd. Huang's original comment -- that he never would have started Nvidia if he knew what he did today -- was grossly taken out of context, he said. In the video, he clarified that he absolutely would start Nvidia again, and that his comment was more about the blissful ignorance of startup founders.

While Zuckerberg's opening comment was just a friendly jab at Huang, it set the tone for Zuckerberg's new attitude toward life and business. The founder of Facebook has spent a lot of time apologizing for Facebook's content moderation issues. But when reflecting on the biggest mistakes of his career, Zuckerberg said his largest one was a "political miscalculation" that he described as a "20-year mistake." Specifically, he said, he'd taken too much ownership for problems allegedly out of Facebook's control. "Some of the things they were asserting that we were doing or were responsible for, I don't actually think we were," said Zuckerberg. "When it's a political problem there are people operating in good faith who are identifying a problem and want something to be fixed, and there are people who are just looking for someone to blame."

An anonymous reader quotes a report from TechCrunch: Meta on Friday published an update on how it plans to comply with the Digital Markets Act (DMA), the European law that aims to promote competition in digital marketplaces, where the law concerns the company's messaging apps, Messenger and WhatsApp. As Meta notes in a blog post, the DMA requires that it provide an option in WhatsApp and Messenger to connect with interoperable third-party messaging services and apps. Meta says it's building notifications into WhatsApp and Messenger to inform users about these third-party integrations and alert them when a newly compatible third-party messaging app comes online. The company also says it's introducing an onboarding flow in WhatsApp and Messenger where users can learn more about third-party chats and switch them on. From the flow, users will be able to set up a designated folder for third-party messages or, alternatively, opt for a combined inbox.

In 2025, Meta will roll out group functionality for third-party chats, and, in 2027, it'll launch voice and video calling in accordance with the DMA. And at some unspecified point in the future, Meta will bring "rich messaging" features for third-party chats to WhatsApp and Messenger, like reactions, direct replies, typing indicators and read receipts, the company says. "We will keep collaborating with third-party messaging services in order to provide the safest and best experience," Meta wrote in the post. "Users will start to see the third-party chat option when a third-party messaging service has built, tested and launched the necessary technology to make the feature a positive and secure user experience."

An anonymous reader quotes a report from Reuters: Facebook owner Meta struck a deal to buy geothermal power from Sage Geosystems to supply its U.S. data centers, it said on Monday, as it races to build out the infrastructure to support its massive investments in energy-hungry artificial intelligence. The first phase of the 150-megawatt project should be operational by 2027 and "significantly" expand the use of geothermal power in the United States, the social media company said. The location has yet to be determined, but the companies said it will be east of the Rocky Mountains. Financial terms of the deal were not disclosed. [...]

Sage, which is based in Houston, is a four-year-old startup developing next-generation technology that it says can be deployed in more locations than traditional geothermal, which requires naturally occurring underground reservoirs of hot water and accounts for 0.4% of U.S. power generation. The company is backed by oil and gas firms Chesapeake Energy and Nabors Industries and venture capital firms Virya and Helium-3 Ventures. The project for Meta would be Sage's largest to date by far. The company said it first validated the technology in the field just two years ago. A Meta spokesperson told Reuters the company expected the Sage Geosystems energy to feed the power grid, rather than directly supplying any specific data center.

According to The Information (paywalled), Meta plans to ship an extremely light mixed reality headset in 2027, codenamed Puffin. It follows a report that the company canceled a high-end headset planned for the same year, which previous reports speculated as being a Quest Pro 2. UploadVR reports: Puffin reportedly resembles "a bulky pair of glasses" and weighs less than 110 grams, yet is an opaque VR-style headset with pancake lenses and passthrough cameras. Its remarkably light weight is apparently being achieved by offloading both the battery and computing hardware to an external tethered puck, which Meta "hopes" will be small enough to fit in the user's pocket.

If the report is accurate, Puffin will be significantly lighter than any other shipping fully functional VR headset to date. For comparison, Meta Quest 3 weighs around 400 grams without its straps and facial interface and around 515 grams with them. Of this weight, the battery is around 70 grams. The report describes Puffin as not including controllers, instead using the gaze-and-pinch input scheme introduced by Apple Vision Pro.

Meta's Twitter rival, Threads, has reached a new milestone of 200 million active users, according to Instagram head Adam Mosseri. "I'm excited to share that we crossed the 200M milestone on @threads," Mosseri wrote. "My hope is that Threads can inspire ideas that bring people together and this amazing community continues to grow." TechCrunch reports: Growth for Threads has been strong. The text-focused social media platform, which launched in July 2023, reached 150 million users in April 2024 and 175 million users in July on its one-year anniversary, before another growth spurt led it to hit 200 million a month later. [...]

Last year, Zuckerberg suggested Threads has a "good chance" of becoming a platform with more than a billion users. On the latest earnings call, the Meta CEO also described the platform as being on a good growth trajectory. "We're making steady progress towards building what looks like it's going to be another major social app. And we are seeing deeper engagement," he said, adding: "I'm quite pleased with the trajectory here."

In the company's second-quarter earnings report on Wednesday, Meta's Reality Labs unit recorded an operating loss of $4.48 billion. CNBC reports: Since late 2020, the Reality Labs unit has generated cumulative losses of about $50 billion, underscoring CEO Mark Zuckerberg's massive investments into the hardware and software that underpins what he says will be the next era of personal computing. Revenue in Reality Labs, largely derived from the company's Quest family of VR headsets and Ray-Ban Meta smart glasses, came in at $353 million, representing growth of 28% from $276 million a year earlier. Analysts were expecting the unit to bring in $371 million.

Meta will pay Texas $1.4 billion to settle a lawsuit alleging the company used personal biometric data without user consent, marking the largest privacy-related settlement ever obtained by a state. The Texas Tribune reports: The 2022 lawsuit, filed by Texas Attorney General Ken Paxton in state court, alleged that Meta had been using facial recognition software on photos uploaded to Facebook without Texans' consent. The settlement will be paid over five years. The attorney general's office did not say whether the money from the settlement would go into the state's general fund or if it would be distributed in some other way. The settlement, announced Tuesday, does not act as an admission of guilt and Meta maintains no wrongdoing. This was the first lawsuit Paxton's office argued under a 2009 state law that protects Texans' biometric data, like fingerprints and facial scans. The law requires businesses to inform and get consent from individuals before collecting such data. It also limits sharing this data, except in certain cases like helping law enforcement or completing financial transactions. Businesses must protect this data and destroy it within a year after it's no longer needed.

In 2011, Meta introduced a feature known as Tag Suggestions to make it easier for users to tag people in their photos. According to Paxton's office, the feature was turned on by default and ran facial recognition on users' photos, automatically capturing data protected by the 2009 law. That system was discontinued in 2021, with Meta saying it deleted over 1 billion people's individual facial recognition data. As part of the settlement, Meta must notify the attorney general's office of anticipated or ongoing activities that may fall under the state's biometric data laws. If Texas objects, the parties have 60 days to attempt to resolve the issue. Meta officials said the settlement will make it easier for the company to discuss the implications and requirements of the state's biometric data laws with the attorney general's office, adding that data protection and privacy are core priorities for the firm.

Thomas Claburn reports via The Register: Meta's machine-learning model for detecting prompt injection attacks -- special prompts to make neural networks behave inappropriately -- is itself vulnerable to, you guessed it, prompt injection attacks. Prompt-Guard-86M, introduced by Meta last week in conjunction with its Llama 3.1 generative model, is intended "to help developers detect and respond to prompt injection and jailbreak inputs," the social network giant said. Large language models (LLMs) are trained with massive amounts of text and other data, and may parrot it on demand, which isn't ideal if the material is dangerous, dubious, or includes personal info. So makers of AI models build filtering mechanisms called "guardrails" to catch queries and responses that may cause harm, such as those revealing sensitive training data on demand, for example. Those using AI models have made it a sport to circumvent guardrails using prompt injection -- inputs designed to make an LLM ignore its internal system prompts that guide its output -- or jailbreaks -- input designed to make a model ignore safeguards. [...]

It turns out Meta's Prompt-Guard-86M classifier model can be asked to "Ignore previous instructions" if you just add spaces between the letters and omit punctuation. Aman Priyanshu, a bug hunter with enterprise AI application security shop Robust Intelligence, recently found the safety bypass when analyzing the embedding weight differences between Meta's Prompt-Guard-86M model and Redmond's base model, microsoft/mdeberta-v3-base. "The bypass involves inserting character-wise spaces between all English alphabet characters in a given prompt," explained Priyanshu in a GitHub Issues post submitted to the Prompt-Guard repo on Thursday. "This simple transformation effectively renders the classifier unable to detect potentially harmful content." "Whatever nasty question you'd like to ask right, all you have to do is remove punctuation and add spaces between every letter," Hyrum Anderson, CTO at Robust Intelligence, told The Register. "It's very simple and it works. And not just a little bit. It went from something like less than 3 percent to nearly a 100 percent attack success rate."

An anonymous reader quotes a report from Ars Technica: The European Commission (EC) has finally taken action to block Meta's heavily criticized plan to charge a subscription fee to users who value privacy on its platforms. Surprisingly, this step wasn't taken under laws like the Digital Services Act (DSA), the Digital Markets Act (DMA), or the General Data Protection Regulation (GDPR). Instead, the EC announced Monday that Meta risked sanctions under EU consumer laws if it could not resolve key concerns about Meta's so-called "pay or consent" model. Meta's model is seemingly problematic, the commission said, because Meta "requested consumers overnight to either subscribe to use Facebook and Instagram against a fee or to consent to Meta's use of their personal data to be shown personalized ads, allowing Meta to make revenue out of it." Because users were given such short notice, they may have been "exposed to undue pressure to choose rapidly between the two models, fearing that they would instantly lose access to their accounts and their network of contacts," the EC said. To protect consumers, the EC joined national consumer protection authorities, sending a letter to Meta requiring the tech giant to propose solutions to resolve the commission's biggest concerns by September 1.

That Meta's "pay or consent" model may be "misleading" is a top concern because it uses the term "free" for ad-based plans, even though Meta "can make revenue from using their personal data to show them personalized ads." It seems that while Meta does not consider giving away personal information to be a cost to users, the EC's commissioner for justice, Didier Reynders, apparently does. "Consumers must not be lured into believing that they would either pay and not be shown any ads anymore, or receive a service for free, when, instead, they would agree that the company used their personal data to make revenue with ads," Reynders said. "EU consumer protection law is clear in this respect. Traders must inform consumers upfront and in a fully transparent manner on how they use their personal data. This is a fundamental right that we will protect." Additionally, the EC is concerned that Meta users might be confused about how "to navigate through different screens in the Facebook/Instagram app or web-version and to click on hyperlinks directing them to different parts of the Terms of Service or Privacy Policy to find out how their preferences, personal data, and user-generated data will be used by Meta to show them personalized ads." They may also find Meta's "imprecise terms and language" confusing, such as Meta referring to "your info" instead of clearly referring to consumers' "personal data." A Meta spokesperson said in a statement: "Subscriptions as an alternative to advertising are a well-established business model across many industries. Subscription for no ads follows the direction of the highest court in Europe and we are confident it complies with European regulation."

Nigeria fined Meta for $220 million on Friday, alleging the tech giant violated the country's local consumer, data protection and privacy laws. Reuters reports: Nigeria's Federal Competition and Consumer Protection Commission (FCCPC) said Meta appropriated the data of Nigerian users on its platforms without their consent, abused its market dominance by forcing exploitative privacy policies on users, and meted out discriminatory and disparate treatment on Nigerians, compared with other jurisdictions with similar regulations. FCCPC chief Adamu Abdullahi said the investigations were jointly held with Nigeria's Data Protection Commission and spanned over 38 months. The investigations found Meta policies don't allow users the option or opportunity to self-determine or withhold consent to the gathering, use, and sharing of personal data, Abdullahi said.

"The totality of the investigation has concluded that Meta over the protracted period of time has engaged in conduct that constituted multiple and repeated, as well as continuing infringements... particularly, but not limited to abusive, and invasive practices against data subjects in Nigeria," Abdullahi said. "Being satisfied with the significant evidence on the record, and that Meta has been provided every opportunity to articulate any position, representations, refutations, explanations or defences of their conduct, the Commission have now entered a final order and issued a penalty against Meta," Abdullahi said. The final order mandates steps and actions Meta must take to comply with local laws, Abdullahi said.

An anonymous reader quotes a report from The Atlantic: Now, after years of contentious relationships with academic researchers, Meta is opening a small pilot program that would allow a handful of them to access Instagram data for up to about six months in order to study the app's effect on the well-being of teens and young adults. The company will announce today that it is seeking proposals that focus on certain research areas -- investigating whether social-media use is associated with different effects in different regions of the world, for example -- and that it plans to accept up to seven submissions. Once approved, researchers will be able to access relevant data from study participants -- how many accounts they follow, for example, or how much they use Instagram and when. Meta has said that certain types of data will be off-limits, such as user-demographic information and the content of media published by users; a full list of eligible data is forthcoming, and it is as yet unclear whether internal information related to ads that are served to users or Instagram's content-sorting algorithm, for example, might be provided. The program is being run in partnership with the Center for Open Science, or COS, a nonprofit. Researchers, not Meta, will be responsible for recruiting the teens, and will be required to get parental consent and take privacy precautions.

According to Axios, Meta will withhold future multimodel AI models from customers in the European Union "due to the unpredictable nature of the European regulatory environment." From the report: Meta plans to incorporate the new multimodal models, which are able to reason across video, audio, images and text, in a wide range of products, including smartphones and its Meta Ray-Ban smart glasses. Meta says its decision also means that European companies will not be able to use the multimodal models even though they are being released under an open license. It could also prevent companies outside of the EU from offering products and services in Europe that make use of the new multimodal models. The company is also planning to release a larger, text-only version of its Llama 3 model soon. That will be made available for customers and companies in the EU, Meta said.

Meta's issue isn't with the still-being-finalized AI Act, but rather with how it can train models using data from European customers while complying with GDPR -- the EU's existing data protection law. Meta announced in May that it planned to use publicly available posts from Facebook and Instagram users to train future models. Meta said it sent more than 2 billion notifications to users in the EU, offering a means for opting out, with training set to begin in June. Meta says it briefed EU regulators months in advance of that public announcement and received only minimal feedback, which it says it addressed. In June -- after announcing its plans publicly -- Meta was ordered to pause the training on EU data. A couple weeks later it received dozens of questions from data privacy regulators from across the region.

The United Kingdom has a nearly identical law to GDPR, but Meta says it isn't seeing the same level of regulatory uncertainty and plans to launch its new model for U.K. users. A Meta representative told Axios that European regulators are taking much longer to interpret existing law than their counterparts in other regions. A Meta representative told Axios that training on European data is key to ensuring its products properly reflect the terminology and culture of the region.

Ahead of its one-year anniversary, Meta CEO Mark Zuckerberg announced that Threads has reached more than 175 million monthly active users. The Verge reports: Back when it arrived in the App Store on July 5th, 2023, Musk was taking a wrecking ball to the service formerly called Twitter and goading Zuckerberg into a literal cage match that never happened. A year later, Threads is still growing at a steady clip -- albeit not as quickly as its huge launch -- while Musk hasn't shared comparable metrics for X since he took over.

As with any social network, and especially for Threads, monthly users only tell part of the growth story. It's telling that, unlike Facebook, WhatsApp, and Instagram, Meta hasn't shared daily user numbers yet. That omission suggests Threads is still getting a lot of flyby traffic from people who have yet to become regular users. I've heard from Meta employees in recent months that much of the app's growth is still coming from it being promoted inside Instagram. Both apps share the same account system, which isn't expected to change.

Brazil's national data protection authority ruled on Tuesday that Meta must stop using data originating in the country to train its artificial intelligence models. The Associated Press reports: Meta's updated privacy policy enables the company to feed people's public posts into its AI systems. That practice will not be permitted in Brazil, however. The decision stems from "the imminent risk of serious and irreparable or difficult-to-repair damage to the fundamental rights of the affected data subjects," the agency said in the nation's official gazette. [...] Hye Jung Han, a Brazil-based researcher for the rights group, said in an email Tuesday that the regulator's action "helps to protect children from worrying that their personal data, shared with friends and family on Meta's platforms, might be used to inflict harm back on them in ways that are impossible to anticipate or guard against."

But the decision regarding Meta will "very likely" encourage other companies to refrain from being transparent in the use of data in the future, said Ronaldo Lemos, of the Institute of Technology and Society of Rio de Janeiro, a think-tank. "Meta was severely punished for being the only one among the Big Tech companies to clearly and in advance notify in its privacy policy that it would use data from its platforms to train artificial intelligence," he said. Compliance must be demonstrated by the company within five working days from the notification of the decision, and the agency established a daily fine of 50,000 reais ($8,820) for failure to do so. In a statement, Meta said the company is "disappointed" by the decision and insists its method "complies with privacy laws and regulations in Brazil."

"This is a step backwards for innovation, competition in AI development and further delays bringing the benefits of AI to people in Brazil," a spokesperson for the company added.

An anonymous reader quotes a report from Ars Technica: Meta continues to hit walls with its heavily scrutinized plan to comply with the European Union's strict online competition law, the Digital Markets Act (DMA), by offering Facebook and Instagram subscriptions as an alternative for privacy-inclined users who want to opt out of ad targeting. Today, the European Commission (EC) announced preliminary findings that Meta's so-called "pay or consent" or "pay or OK" model -- which gives users a choice to either pay for access to its platforms or give consent to collect user data to target ads -- is not compliant with the DMA. According to the EC, Meta's advertising model violates the DMA in two ways. First, it "does not allow users to opt for a service that uses less of their personal data but is otherwise equivalent to the 'personalized ads-based service." And second, it "does not allow users to exercise their right to freely consent to the combination of their personal data," the press release said.

Now, Meta will have a chance to review the EC's evidence and defend its policy, with today's findings kicking off a process that will take months. The EC's investigation is expected to conclude next March. Thierry Breton, the commissioner for the internal market, said in the press release that the preliminary findings represent "another important step" to ensure Meta's full compliance with the DMA. "The DMA is there to give back to the users the power to decide how their data is used and ensure innovative companies can compete on equal footing with tech giants on data access," Breton said. A Meta spokesperson told Ars that Meta plans to fight the findings -- which could trigger fines up to 10 percent of the company's worldwide turnover, as well as fines up to 20 percent for repeat infringement if Meta loses. The EC agreed that more talks were needed, writing in the press release, "the Commission continues its constructive engagement with Meta to identify a satisfactory path towards effective compliance." Meta continues to claim that its "subscription for no ads" model was "endorsed" by the highest court in Europe, the Court of Justice of the European Union (CJEU), last year.

"Subscription for no ads follows the direction of the highest court in Europe and complies with the DMA," Meta's spokesperson said. "We look forward to further constructive dialogue with the European Commission to bring this investigation to a close."

Meta rolled out its ad-free subscription service option last November. "Depending on where you purchase it will cost $10.5/month on the web or $13.75/month on iOS and Android," said the company in a blog post. "Regardless of where you purchase, the subscription will apply to all linked Facebook and Instagram accounts in a user's Accounts Center. As is the case for many online subscriptions, the iOS and Android pricing take into account the fees that Apple and Google charge through respective purchasing policies."

Since May, Meta has been labeling photos created with AI tools on its social networks to help users better identify the content they're consuming. However, as TechCrunch's Ivan Mehta reports, this approach has faced criticism as many photos not created using AI tools have been incorrectly labeled, prompting Meta to reevaluate its labeling strategy to better reflect the actual use of AI in images. From the report: There are plenty of examples of Meta automatically attaching the label to photos that were not created through AI. For example, this photo of Kolkata Knight Riders winning the Indian Premier League Cricket tournament. Notably, the label is only visible on the mobile apps and not on the web. Plenty of other photographers have raised concerns over their images having been wrongly tagged with the "Made with AI" label. Their point is that simply editing a photo with a tool should not be subject to the label.

Former White House photographer Pete Souza said in an Instagram post that one of his photos was tagged with the new label. Souza told TechCrunch in an email that Adobe changed how its cropping tool works and you have to "flatten the image" before saving it as a JPEG image. He suspects that this action has triggered Meta's algorithm to attach this label. "What's annoying is that the post forced me to include the 'Made with AI' even though I unchecked it," Souza told TechCrunch.

Meta would not answer on the record to TechCrunch's questions about Souza's experience or other photographers' posts who said their posts were incorrectly tagged. However, after publishing of the story, Meta said the company is evaluating its approach to indicate labels reflect the amount of AI used in an image. "Our intent has always been to help people know when they see content that has been made with AI. We are taking into account recent feedback and continue to evaluate our approach so that our labels reflect the amount of AI used in an image," a Meta spokesperson told TechCrunch. "For now, Meta provides no separate labels to indicate if a photographer used a tool to clean up their photo, or used AI to create it," notes TechCrunch. "For users, it might be hard to understand how much AI was involved in a photo."

"Meta's label specifies that 'Generative AI may have been used to create or edit content in this post' -- but only if you tap on the label. Despite this approach, there are plenty of photos on Meta's platforms that are clearly AI-generated, and Meta's algorithm hasn't labeled them."

Meta has released the Threads API for developers to build "unique integrations" into the text-based conversation app. The move could potentially result in third-party apps. The Verge reports: "People can now publish posts via the API, fetch their own content, and leverage our reply management capabilities to set reply and quote controls, retrieve replies to their posts, hide, unhide or respond to specific replies," explains Jesse Chen, director of engineering at Threads.

Chen says that insights into Threads posts are "one of our top requested features for the API," so Meta is allowing developers to see the number of views, likes, replies, reposts, and quotes on Threads posts through the API. Meta has published plenty of documentation about how developers can get started with the Threads API, and there's even an open-source Threads API sample app on GitHub.

Thomas Claburn reports via The Register: Meta allegedly tried to discredit university researchers in Brazil who had flagged fraudulent adverts on the social network's ad platform. Nucleo, a Brazil-based news organization, said it has obtained government documents showing that attorneys representing Meta questioned the credibility of researchers from NetLab, which is part of the Federal University of Rio de Janeiro (UFRJ). NetLab's research into Meta's ads contributed to Brazil's National Consumer Secretariat (Senacon) decision in 2023 to fine Meta $1.7 million (9.3 million BRL), which is still being appealed. Meta (then Facebook) was separately fined of $1.2 million (6.6 million BRL) related to Cambridge Analytica.

As noted by Nucleo, NetLab's report showed that Facebook, despite being notified about the issues, had failed to remove more than 1,800 scam ads that fraudulently used the name of a government program that was supposed to assist those in debt. In response to the fine, attorneys representing Meta from law firm TozziniFreire allegedly accused the NetLab team of bias and of failing to involve Meta in the research process. Nucleo says that it obtained the administrative filing through freedom of information requests to Senacon. The documents are said to date from December 26 last year and to be part of the ongoing case against Meta. A spokesperson for NetLab, who asked not to be identified by name due to online harassment directed at the organization's members, told The Register that the research group was aware of the Nucleo report. "We were kind of surprised to see the account of our work in this law firm document," the spokesperson said. "We expected to be treated with more fairness for our work. Honestly, it comes at a very bad moment because NetLab particularly, but also Brazilian science in general, is being attacked by far-right groups."

On Thursday, more than 70 civil society groups including NetLab published an open letter decrying Meta's legal tactics. "This is an attack on scientific research work, and attempts at intimidation of researchers and researchers who are performing excellent work in the production of knowledge from empirical analysis that have been fundamental to qualify the public debate on the accountability of social media platforms operating in the country, especially with regard to paid content that causes harm to consumers of these platforms and that threaten the future of our democracy," the letter says. "This kind of attack and intimidation is made even more dangerous by aligning with arguments that, without any evidence, have been used by the far right to discredit the most diverse scientific productions, including NetLab itself." The claim, allegedly made by Meta's attorneys, is that the ad biz was "not given the opportunity to appoint a technical assistant and present questions" in the preparation of the NetLabs report. This is particularly striking given Meta's efforts to limit research into its ad platform. A Meta spokesperson told The Register: "We value input from civil society organizations and academic institutions for the context they provide as we constantly work toward improving our services. Meta's defense filed with the Brazilian Consumer Regulator questioned the use of the NetLab report as legal evidence, since it was produced without giving us prior opportunity to contribute meaningfully, in violation of local legal requirements."

Some artists are jumping ship for the anti-AI portfolio app Cara after Meta began using Instagram content to train its AI models. Fast Company explains: The portfolio app bills itself as a platform that protects artists' images from being used to train AI, and only allowing AI content to be posted if it's clearly labeled. Based on the number of new users the Cara app has garnered over the past few days, there seems to be a need. Between May 31 and June 2, Cara's user base tripled from less than 100,000 to more than 300,000 profiles, skyrocketing to the top of the app store. [...] Cara is a social networking app for creatives, in which users can post images of their artwork, memes, or just their own text-based musings. It shares similarities with major social platforms like X (formerly Twitter) and Instagram on a few fronts. Users can access Cara through a mobile app or on a browser. Both options are free to use. The UI itself is like an arts-centric combination of X and Instagram. In fact, some UI elements seem like they were pulled directly from other social media sites. (It's not the most innovative approach, but it is strategic: as a new app, any barriers to potential adoption need to be low).

Cara doesn't train any AI models on its content, nor does it allow third parties to do so. According to Cara's FAQ page, the app aims to protect its users from AI scraping by automatically implementing "NoAI" tags on all of its posts. The website says these tags "are intended to tell AI scrapers not to scrape from Cara." Ultimately, they appear to be html metadata tags that politely ask bad actors not to get up to any funny business, and it's pretty unlikely that they hold any actual legal weight. Cara admits as much, too, warning its users that the tags aren't a "fully comprehensive solution and won't completely prevent dedicated scrapers." With that in mind, Cara assesses the "NoAI" tagging system as a "a necessary first step in building a space that is actually welcoming to artists -- one that respects them as creators and doesn't opt their work into unethical AI scraping without their consent."

In December, Cara launched another tool called Cara Glaze to defend its artists' work against scrapers. (Users can only use it a select number of times.) Glaze, developed by the SAND Lab at University of Chicago, makes it much more difficult for AI models to accurately understand and mimic an artist's personal style. The tool works by learning how AI bots perceive artwork, and then making a set of minimal changes that are invisible to the human eye but confusing to the AI model. The AI bot then has trouble "translating" the art style and generates warped recreations. In the future, Cara also plans to implement Nightshade, another University of Chicago software that helps protect artwork against AI scapers. Nightshade "poisons" AI training data by adding invisible pixels to artwork that can cause AI software to completely misunderstand the image. Beyond establishing shields against data mining, Cara also uses a third party service to detect and moderate any AI artwork that's posted to the site. Non-human artwork is forbidden, unless it's been properly labeled by the poster.

An anonymous reader quotes a report from Quartz: Mark Zuckerberg has assembled some of his fellow tech chiefs into an advisory council to guide Meta on its artificial intelligence and product developments. The Meta Advisory Group will periodically meet with Meta's management team, Bloomberg reported. Its members include: Stripe CEO and co-founder Patrick Collison, former GitHub CEO Nat Friedman, Shopify CEO Tobi Lutke, and former Microsoft executive and investor Charlie Songhurst.

"I've come to deeply respect this group of people and their achievements in their respective areas, and I'm grateful that they're willing to share their perspectives with Meta at such an important time as we take on new opportunities with AI and the metaverse," Zuckerberg wrote in an internal note to Meta employees, according to Bloomberg. The advisory council differs from Meta's 11-person board of directors because its members are not elected by shareholders, nor do they have fiduciary duty to Meta, a Meta spokesperson told Bloomberg. The spokesperson said that the men will not be paid for their roles on the advisory council. TechCrunch notes that the council features "only white men on it." This "differs from Meta's actual board of directors and its Oversight Board, which is more diverse in gender and racial representation," reports TechCrunch.

"It's telling that the AI advisory council is composed entirely of businesspeople and entrepreneurs, not ethicists or anyone with an academic or deep research background. ... it's been proven time and time again that AI isn't like other products. It's a risky business, and the consequences of getting it wrong can be far-reaching, particularly for marginalized groups."