Tags: deepfake

Meta has taken down an AI‑generated deepfake video that falsely portrayed independent presidential candidate Catherine Connolly announcing her withdrawal from the race. The video, posted by an account named RTÉ News AI, was shared nearly 30,000 times on Facebook before removal. Connolly condemned the clip as a "disgraceful attempt to mislead voters" and affirmed her continued candidacy. Meta cited violations of its community standards on impersonation, while Irish media regulator Coimisiún na Meán confirmed the platform’s swift response. The incident highlights ongoing challenges in policing political deepfakes on social media. Lire la suite

OpenAI has released Sora, an AI video generator that creates high‑resolution, synchronized videos from text prompts. The tool includes a moving watermark, built‑in metadata, and a "cameo" feature that can insert real‑world likenesses into generated scenes. While Sora’s capabilities are praised for creativity and ease of use, experts warn it could simplify the production of deepfakes and misinformation. Platforms such as Meta, TikTok, and YouTube are experimenting with AI‑content labeling, and tools like the Content Authenticity Initiative’s verifier can help identify Sora‑generated media. The debate highlights the tension between innovation and the need for robust safeguards. Lire la suite

YouTube has introduced a new AI‑powered detection feature that lets creators in its Partner Program locate and flag videos that use their likeness without permission. After confirming their identity, creators can review flagged content in a new Content Detection tab within YouTube Studio and request removal of unauthorized AI‑generated videos. The rollout begins with an initial group of creators and will expand over the coming months. The tool, similar to Content ID, first appeared in a pilot with talent represented by Creative Artists Agency and joins other recent policies that require labeling of AI‑generated material and restrict AI‑mimicked music. Lire la suite

YouTube is rolling out a beta likeness detection tool that aims to identify AI‑generated videos that misuse a creator’s face. The feature, similar to the platform’s copyright detection system, requires creators to verify their identity with government ID and a facial video. Initially limited to a small group, the tool is now being offered to more eligible creators, giving them a way to protect their likeness from synthetic content that could spread misinformation or damage their brand. Lire la suite

YouTube is beginning to roll out a new likeness detection feature aimed at identifying and removing videos that alter a person’s facial appearance with artificial intelligence. The tool, initially available to members of the YouTube Partner Program, requires creators to submit a government ID and a short selfie for verification. It works similarly to Content ID by scanning uploads for facial matches, allowing individuals to flag infringing content for removal. The system currently focuses only on visual modifications and does not address AI‑altered voices. Lire la suite

OpenAI has halted the ability of its Sora video‑generation tool to create videos featuring Martin Luther King Jr. following a request from King Inc., the estate that manages the civil‑rights leader’s legacy. The company said it worked with the estate to address how King’s likeness is represented after users produced disrespectful depictions, including videos that showed King making monkey noises and wrestling with Malcolm X. Families of other public figures, such as Zelda Williams and Bernice A. King, also voiced concerns about AI‑generated content. OpenAI emphasized the importance of free‑speech interests while asserting that families should control the use of their relatives’ likenesses, and it noted that other estates can request similar restrictions. Lire la suite

OpenAI has launched Sora, an AI‑driven video generation app that lets users create short clips from text prompts. The tool, powered by the company’s latest model, has quickly attracted a large user base, surpassing a million downloads in its first week. While experts praise Sora’s potential to spark a new wave of digital creativity, they also warn of its capacity for deep‑fake misuse and the broader impact on social media dynamics. The app’s design—ten‑second videos, infinite scroll, and a focus on AI‑generated content rather than personal uploads—has sparked debate about authenticity, addiction, and the future of online interaction. Lire la suite

Google announced the launch of Veo 3.1, the latest version of its AI video generator, now available to paying Gemini users through Flow, the Gemini API, and Vertex AI. The update adds asset‑combining tools, object insertion, extended clip length, and AI‑driven transitions between still images. The move comes as OpenAI promotes its Sora video model, sparking debate over deep‑fake risks and copyright challenges in generative media. Lire la suite

OpenAI introduced Sora 2 as a video‑generation model that blocks depictions of public figures by default, yet a loophole permits creators to insert deceased personalities into AI‑generated clips. The platform’s moving watermark aims to signal synthetic content, but families and fans are upset by the posthumous portrayals, exemplified by Zelda Williams’ public plea to stop receiving AI videos of her late father. OpenAI also offers a “cameo” feature that lets living individuals voluntarily upload their likenesses, with controls to revoke access and moderate usage, highlighting the tension between creative freedom and ethical safeguards. Lire la suite

OpenAI announced updates to its invite‑only Sora app that give users more control over how their likenesses are used in AI‑generated videos. The changes let creators add keyword or scenario restrictions to their cameo profiles and make the watermark on downloaded videos more visible. The move follows criticism from copyright experts and recent legal challenges involving AI‑generated content. OpenAI’s head of Sora, Bill Peebles, said the company is also refining safety guardrails to reduce false negatives while addressing concerns from creators and rights holders. Lire la suite

OpenAI introduced Sora, a social media app built around AI‑generated videos. The platform lets users scroll an endless feed, like, comment, and share clips created with the upgraded Sora 2 model. A standout feature is “Cameos,” which allows creators to upload their likeness and permit others to generate deep‑fake style videos featuring them. The app includes safeguards such as watermarks and metadata tags that indicate AI origins, yet it raises concerns about the ease of producing realistic deepfakes, potential misuse, and the environmental impact of generating high‑quality video content. Lire la suite

OpenAI has released Sora, an iOS app that lets users create short AI‑generated videos featuring their own digital likenesses. The platform offers a scrollable feed of bite‑size clips and includes built‑in safety guardrails to restrict sexual content, graphic violence, extremist propaganda, hate speech, and self‑harm. Users can control who may use their likeness and can see any details about generated videos that involve them. While the app showcases impressive realism, OpenAI acknowledges the potential for misuse and has implemented multiple safeguards. Lire la suite

OpenAI has launched Sora, a new social platform that lets users generate AI‑driven videos featuring realistic likenesses of public figures, including CEO Sam Altman. The app’s feed is flooded with user‑created clips that place Altman in surreal scenarios, prompting discussions about copyright, consent, and the potential for misuse. While OpenAI touts safety controls and user‑managed cameo settings, critics warn that the technology could enable disinformation and raise ethical concerns about deepfake creation. Lire la suite

President Donald Trump posted a short AI‑generated video to his Truth Social account that depicted a deepfake version of Senator Chuck Schumer delivering crude insults while Representative Hakeem Jeffries appeared in the background. The synthetic clip, lasting about 35 seconds, included vulgar language, racial overtones and bizarre conspiracy references. The video targeted the two Democratic leaders during negotiations over a government funding agreement. The incident highlights growing concerns about political deepfakes and the misuse of artificial‑intelligence tools to spread misinformation and inflame partisan tensions. Lire la suite

President Donald Trump briefly shared on Truth Social a video that appeared to be a Fox News segment featuring Lara Trump announcing a supposed MedBed hospital and national MedBed card system—both of which do not exist. The clip was later removed. Fox News confirmed the segment never aired, and observers noted the video was fully AI‑generated, including Trump’s dialogue in the Oval Office. MedBeds are an imagined medical device touted by conspiracy circles as a cure‑all. While some followers recognized the AI origin, many still interpreted the post as confirmation of the device’s reality. The White House is likely to downplay the incident. Lire la suite

Spotify has rolled out a suite of policies aimed at curbing AI‑generated music that impersonates real artists without permission. The new rules require explicit artist consent for any AI‑replicated vocals and mandate that AI usage be disclosed in track credits. Alongside this, the platform is deploying an AI‑aware spam filter to target low‑effort, algorithm‑gaming uploads, which it says removed more than 75 million spammy tracks in the past year. Spotify also plans to offer nuanced metadata so listeners can see exactly how much AI contributed to a song, signaling a move toward greater transparency in the streaming ecosystem. Lire la suite

Google announced a partnership with the UK nonprofit StopNCII to expand its defenses against non‑consensual intimate imagery (NCII), commonly known as revenge porn. The collaboration will see Google incorporate StopNCII's hash‑based system, allowing user‑generated digital fingerprints to block unwanted intimate content from appearing in search results. The service protects privacy by never uploading the original image, and it works alongside other platforms already partnered with StopNCII. While the system is not a complete solution, it marks a significant step for Google in reducing the burden on victims of NCII. Lire la suite

A spoof website called the Center for the Alignment of AI Alignment Centers presents itself as a hub for AI safety researchers, but hides jokes and absurd claims throughout. Created by the team behind a wearable “Box” for women, the site features calming visuals that conceal hidden messages, a job board that invites applicants who believe AGI will end humanity in weeks, and a generative AI tool that promises to build a fake AI center in minutes. Researchers such as Kendra Albert initially mistook it for a legitimate effort before recognizing the satire. Lire la suite

Vodafone released a commercial featuring an AI-generated avatar as the spokesperson, describing the effort as an experiment in using artificial intelligence for advertising. The company said it is testing different styles of advertising with AI, noting the technology’s growing presence in everyday life. This follows a previous AI-generated ad from last year that attracted controversy, and reflects a broader trend of brands exploring virtual influencers and deepfake technology. Lire la suite

AI-driven impersonation scams are exploding, using voice cloning and deepfake video to mimic trusted individuals. Criminals target victims through phone calls, video meetings, messages, and emails, often creating urgent requests for money or confidential information. Experts advise slowing down, verifying identities, and adding multi‑factor authentication to protect against these sophisticated attacks. The rise is driven by improved technology, lower costs, and broader accessibility, affecting both consumers and corporations. Lire la suite