tldrdailynews.com logo
Tag

Watermark

All articles tagged with #watermark

Open-Source AI Watermark Remover Goes Viral, Sparking Debate Over AI Attribution
technology20 days ago

Open-Source AI Watermark Remover Goes Viral, Sparking Debate Over AI Attribution

Guillaume Meyer’s open-source project Watermarks Remover, created in response to Anthropic’s announcement of invisible AI watermarks, went viral after a few hours of work and now has contributed contributors and millions of impressions. Meyer argues the watermark approach can cause false positives and mislabel non-AI content (like Grammarly-generated text), highlighting regulatory and trust concerns. The tool demonstrates a broader tension between attributions, AI regulation, and practical use, and Meyer says he may turn it into a business while plans to adapt as detectors evolve.

via Business Insider|
#ai#open-source#regulation
Neverness to Everness Faces Fresh AI Art Scandal Over Watermarked Seahorse
technology25 days ago

Neverness to Everness Faces Fresh AI Art Scandal Over Watermarked Seahorse

Neverness to Everness is again at the center of a genAI controversy after players in China spotted a watermark from Doubao AI on a seahorse artwork in the 1.3 update, released to China a week early ahead of the global rollout. Hotta Games had previously said AI assets were limited to backgrounds and not to characters or stories, and this latest slip renews criticism of hidden AI-generated art in the game.

via Kotaku|
#ai#artwork#hotta-games
Google Gemini Lets Users Remove Visible AI Watermarks, Sparking Detection Concerns
technology29 days ago

Google Gemini Lets Users Remove Visible AI Watermarks, Sparking Detection Concerns

Google Gemini will soon let users choose whether AI-generated images, videos, and music carry visible watermarks. It relies on C2PA metadata and an invisible SynthID watermark for verification, but metadata can be stripped and SynthID has shown resilience to some edits yet can be defeated by adversarial techniques and existing bypass tools. The move raises concerns about the erosion of our ability to recognize AI-generated content, with rollout excluding countries where watermarking is legally required.

via gizmodo.com|
#c2pa#google#google-gemini
Google Lets Users Toggle Off Visible AI Watermarks
technology29 days ago

Google Lets Users Toggle Off Visible AI Watermarks

Google announced that users will be able to disable visible watermarks on AI-generated images and videos with a toggle for the Nano Banana, Omni and Lyria models, with Gemini and Flow to follow. The invisible SynthID watermark and C2PA metadata remain, and there will be no toggles for those. The rollout will exclude regions where laws require watermarks (notably the EU). Google says the change balances creative control and safety as AI labeling laws expand, alongside other firms pursuing transparency tools.

via Engadget|
#ai#google#synthid
Claude's invisible watermarks aim to prove AI processing under EU AI Act
technology1 month ago

Claude's invisible watermarks aim to prove AI processing under EU AI Act

Anthropic will embed invisible watermarks in Claude outputs to comply with the EU AI Act, applying to new models globally and to processed content where supported, with a forthcoming detection API; but the marks can be bypassed, may mislabel human edits as AI-processed, and their effectiveness remains uncertain, with penalties up to €15 million or 3% of global revenue for noncompliance.

via arstechnica.com|
#ai-regulation#anthropic#claude
Anthropic to watermark Claude outputs invisibly as industry fights AI slop
technology1 month ago

Anthropic to watermark Claude outputs invisibly as industry fights AI slop

Anthropic plans to embed an imperceptible watermark in Claude-generated text (and images) to trace AI output across copying and editing, aligning with the EU AI Act’s transparency rules; it’s part of a broader industry push to police AI-generated content and curb “AI slop,” with platforms like Substack and YouTube adopting labeling and safeguards. Watermarks face limitations and can be removed or bypassed, but the move signals rising pressure to distinguish human from machine writing.

via Fortune|
#ai-regulation#anthropic#claude
Google Earth scrubs AI imagery feature after 24-hour rollout amid misinformation concerns
technology1 month ago

Google Earth scrubs AI imagery feature after 24-hour rollout amid misinformation concerns

Google briefly added an AI image generator called Nano Banana to Google Earth and then rolled it back within a day amid concerns it could be used to spread misinformation. Google said generated images would include a SynthID watermark and could be verified via Gemini or Lens, and that the feature was removed to implement stronger guardrails, despite initial assurances that watermarked AI outputs wouldn’t appear in the main Earth view.

via Mashable|
#ai-generated-imagery#google-earth#misinformation
Google Earth's AI image editor pulled after a day amid deepfake worries
tech1 month ago

Google Earth's AI image editor pulled after a day amid deepfake worries

Google shut down the AI image-editing feature in Google Earth just one day after its launch, citing misuse and policy violations as users generated AI deepfakes (including refugee and war-related imagery) that could spread; the company said watermarked outputs didn’t appear in the main Earth experience and will roll back the tool while adding stronger guardrails, as researchers demonstrated that even with watermarking a detector could be bypassed.

via The Verge|
#ai-artificial-intelligence#deepfake#google-earth
EU Debuts AI-Labeling Rules for Content Across Platforms
technology1 month ago

EU Debuts AI-Labeling Rules for Content Across Platforms

EU's new AI transparency rules require clear labeling of AI-generated or substantially manipulated content, including visible labels and technical markers like watermarks; applies mainly to professional content with exemptions for personal use and artistic works, and requires persistent detection across platforms; compliance by Dec 2 with penalties for noncompliance; platforms and tech giants like Meta, TikTok, and Google are adopting labeling practices and standards.

via TechSpot|
#ai#ai-generated-content#eu-regulation
YouTube Shorts Debuts AI Avatars for Self-Cloned Videos
technology5 months ago

YouTube Shorts Debuts AI Avatars for Self-Cloned Videos

YouTube is rolling out an AI avatar feature for Shorts that lets creators record a live selfie to generate a digital version of themselves, which can be used in new clips or added to eligible Shorts; avatars must be used in the creator's own original videos, with remix controls and automatic deletion after three years if unused; all avatar content will be clearly labeled as AI-generated with watermarking like SynthID and C2PA; access is gradual, requiring 18+, an existing channel, and is part of Google's Gemini AI toolkit; it comes amid concerns about deepfakes and other AI-generated content.

via The Verge|
#ai#avatar#deepfake
OpenAI's Sora 2 Faces Copyright Concerns Amid Viral Success
technology11 months ago

OpenAI's Sora 2 Faces Copyright Concerns Amid Viral Success

OpenAI's Sora app faced unexpected copyright issues and public outcry over AI-generated content, leading to a policy reversal where rightsholders can now specify how their likenesses are used. The company is working on additional controls and clearer watermarks amid concerns about misinformation and misuse, while also expanding access to its AI video technology and infrastructure investments.

via The Verge|
#ai-generated-videos#copyright#openai
"Samsung's Galaxy S24 AI: Removing Watermarks and Dominating the Smartphone Race"
technology2 years ago

"Samsung's Galaxy S24 AI: Removing Watermarks and Dominating the Smartphone Race"

Samsung's new Galaxy S24, equipped with AI capabilities, features a watermark intended to indicate if an image was AI-generated, but the phone's own AI-based object erase tool can easily remove the watermark. This oversight raises concerns about the spread of AI-altered or fake images online, as watermarks and metadata can be easily manipulated. While big AI companies advocate for watermarking as a solution, the issue highlights the ongoing challenge of ensuring the authenticity of digital images in the age of advanced AI technology.

via Gizmodo|
#ai#deepfake#image-manipulation