Adobe, OpenAI and Microsoft are among the major firms backing a California bill that would require tech companies to label AI-generated content with watermarks embedded in the metadata. Such data is easily accessible via browser for material circulated on the Internet, and the initiative would likely involve a campaign to educate the general public on how to find it. The proposed law encompasses video and audio as well as images. The three companies currently supporting the bill initially opposed it, using terms like “unworkable” and “overly burdensome.” Continue reading Bill Mandating GenAI Watermarks Gains Support in California

Grok-2 and Grok-2 mini, the latest generative chatbots from Elon Musk’s xAI, create images with seemingly few guardrails. Early pictures of notable personalities such as Bill Gates, Donald Trump and Kamala Harris in questionable or compromising settings may not appear photorealistic to a trained eye, but they are still described in many cases to be quite realistic. Powered by the FLUX.1 AI model from Black Forest Labs, Grok-2 and Grok-2 mini are available in beta on X social for Premium and Premium+ subscribers and will be coming to xAI’s enterprise API later this month, according to the company. Continue reading xAI’s Grok-2 Generates Realistic Images with Few Guardrails

YouTube, which began testing crowdsourced fact-checking in June, is now expanding the experiment by inviting users to try the feature. Likened to the Community Notes accountability method introduced by Twitter and continued under X, YouTube’s as yet unnamed feature lets users provide context and corrections to posts that might be misleading or false. “You can sign up to submit notes on videos you find inaccurate or unclear,” YouTube explains, adding that “after submission, your note is reviewed and rated by others.” Notes widely rated as helpful “may be published and appear below the video.” Continue reading YouTube Tests Expanded Community Fact-Checking for Video

Meta Platforms has expanded its licensing deal with Universal Music Group, which now covers users sharing songs from the UMG library across Facebook, Instagram, Threads, WhatsApp and Horizon without fear of copyright violation. As part of a multiyear partnership, Meta and UMG will work together to address “unauthorized AI-generated content that could affect artists and songwriters.” This could encompass everything from users experimenting with voice copying or scraping songs for AI mashups to enterprise level crawling and theft for model training, and suggests Meta will be implementing a filtering system to detect infractions. Continue reading Meta, UMG Music Deal Emphasizes the Responsible Use of AI

The U.S. Copyright Office is warning of an urgent national need for protection against deepfakes. In the first installment of a multipart report on the adverse effects of artificial intelligence on copyright, the office recommends the immediate enactment of a law to combat AI-driven “digital replicas.” Acknowledging that copyright has always had a symbiotic relationship with technology, as well as AI’s tremendous potential, the report nonetheless decries the proliferation of AI-generated deepfakes, “from celebrities’ images endorsing products to politicians’ likenesses seeking to affect voter behavior.” Continue reading Copyright Office Calls for Federal Law Regulating Deepfakes

The Senate has introduced the NO FAKES Act (Nurture Originals, Foster Art, and Keep Entertainment Safe) to protect artists — their voices and visual likenesses — from the proliferation of deepfakes and digital replicas created without consent. The bipartisan bill seeks to impose liability for damages to those who violate the proposed new law. If passed, the NO FAKES Act would be the first federal protection from AI image appropriation, supporters say. Those who’ve rallied to the cause include SAG-AFTRA, the Recording Industry Association of America, the Motion Picture Association, Disney and major talent agencies. Continue reading Senate Introduces NO FAKES Act to Address Deepfakes and AI

The Senate has introduced a bill that takes on tamping down deepfakes while also protecting creative content from use for AI model training. The Content Origin Protection and Integrity from Edited and Deepfaked Media Act, to be known as the COPIED Act, seeks to enact safeguards to protect journalists, actors, songwriters and other artists “against AI-driven theft,” while establishing new federal transparency guidelines for marking, authenticating and detecting AI-generated content. Emphasizing accountability, the bill stipulates that those found in violation will be subject to legal action. Continue reading COPIED Act Seeks to Protect from Deepfakes, Training Abuse

Deepfake videos are becoming increasingly problematic, not only in spreading disinformation on social media but also in enterprise attacks. Now researchers at Drexel University College of Engineering say they have developed an advanced algorithm with a 98 percent accuracy rate in detecting deepfake videos. Called the MISLnet algorithm, for the school’s Multimedia and Information Security Lab where it was invented, the platform uses machine learning to recognize and extract the “digital fingerprints” of video generators including Stable Video Diffusion, VideoCrafter and CogVideo. Continue reading Drexel Claims Its AI Has 98 Percent Rate Detecting Deepfakes

Meta Platforms is publicly releasing five new AI models from its Fundamental AI Research (FAIR) team, which has been experimenting with artificial intelligence since 2013. These models including image-to-text, text-to-music generation, and multi-token prediction tools. Meta is introducing a new technique called AudioSeal, an audio watermarking technique designed for the localized detection of AI-generated speech. “AudioSeal makes it possible to pinpoint AI-generated segments within a longer audio snippet,” according to Meta. The feature is timely in light of concern about potential misinformation surrounding the fall presidential election. Continue reading Meta’s FAIR Team Announces a New Collection of AI Models

YouTube is experimenting with a feature that allows viewers to add contextual “Notes” under videos, similar to what X does with its Community Notes. The Google-owned company says the intent is to provide clarity around things like “when a song is meant to be a parody,” when newly reviewed products are available for purchase, or “when older footage is mistakenly portrayed as a current event.” However, the timing preceding a pivotal U.S. presidential election and facing concerns about deepfakes and  misinformation is no doubt intentional. The pilot will initially be available on mobile in the United States. Continue reading YouTube to Tackle Misinformation with Crowdsourced Notes

London-based AI-startup Synthesia, which creates avatars for enterprise-level generative video presentations, has added “Expressive Avatars” to its feature kit. Powered by Synthesia’s new Express-1 model, these fourth-generation avatars have achieved a new benchmark in realism by using contextual expressions that approximates human emotion, the company says. Express-1 has been trained “to understand the intricate relationship between what we say and how we say it,” allowing Expressive Avatars to perform a script with the correct vocal tone, body language and lip movement, “like a real actor,” according to Synthesia. Continue reading Synthesia Express-1 Model Gives ‘Expressive Avatars’ Emotion

Microsoft has developed VASA, a framework for generating lifelike virtual characters with vocal capabilities including speaking and singing. The premiere model, VASA-1, can perform the feat in real time from a single static image and a vocalization clip. The research demo showcases realistic audio-enhanced faces that can be fine-tuned to look in different directions or change expression in video clips of up to one minute at 512 x 512 pixels and up to 40fps “with negligible starting latency,” according to Microsoft, which says “it paves the way for real-time engagements with lifelike avatars that emulate human conversational behaviors.” Continue reading Microsoft’s VASA-1 Can Generate Talking Faces in Real Time

OpenAI has debuted a new text-to-voice generation platform called Voice Engine, available in limited access. Voice Engine can generate a synthetic voice from a 15-second clip of someone’s voice. The synthetic voice can then read a provided text, even translating to other languages. For now, only a handful of companies are using the tech under a strict usage policy as OpenAI grapples with the potential for misuse. “These small scale deployments are helping to inform our approach, safeguards, and thinking about how Voice Engine could be used for good across various industries,” OpenAI explained. Continue reading OpenAI Voice Cloning Tool Needs Only a 15-Second Sample

YouTube has added new rules requiring those uploading realistic-looking videos that are “made with altered or synthetic media, including generative AI” to label them using a new tool in Creator Studio. The new labeling “is meant to strengthen transparency with viewers and build trust between creators and their audience,” YouTube says, listing examples of content that require disclosure as “likeness of a realistic person” including voice as well as image, “altering footage of real events or places” and “generating realistic scenes” of fictional major events, “like a tornado moving toward a real town.” Continue reading YouTube Adds GenAI Labeling Requirement for Realistic Video

Artificial intelligence imaging service Midjourney has been embraced by storytellers who have also been clamoring for a feature that enables characters to regenerate consistently across new requests. Now Midjourney is delivering that functionality with the addition of the new “–cref” tag (short for Character Reference), available for those who are using Midjourney v6 on the Discord server. Users can achieve the effect by adding the tag to the end of text prompts, followed by a URL that contains the master image subsequent generations should match. Midjourney will then attempt to repeat the particulars of a character’s face, body and clothing characteristics. Continue reading Midjourney Creates a Feature to Advance Image Consistency