Automated trust-and-safety at scale: classifiers, deepfake detection and policy enforcement.
266 articles
The Indian government is summoning Meta executives following a BBC investigation that revealed Instagram was running paid ads promoting child sexual abuse material. The Ministry of Electronics and Information Technology is demanding a formal explanation.
The Next Web (TNW) · 77d ago
India's technology minister has ordered officials to summon Meta Platforms Inc. regarding the prevalence of child sexual abuse content on Instagram.
Bloomberg Technology · 78d ago
A new report reveals that safety features on popular social media apps are frequently ineffective. Teens are still able to access harmful content, interact with strangers, and bypass parental time restrictions.
New York Times Technology · 82d ago
Yahoo Japan has introduced a new feature that uses OpenAI's API to analyze and categorize user comments. The system visualizes the main discussion points of Yahoo Comments using AI-generated graphs.
ITmedia AI+ (日本) · 82d ago
Apple Books is struggling with a proliferation of AI-generated knockoff versions of existing books. This mirrors similar content moderation challenges currently faced by Amazon.
Digital Trends · 84d ago
Rakuten Kobo rejected nearly half of all submissions to its self-publishing platform in 2025. CEO Michael Tamblyn stated that over 80% of these rejections were due to low-quality, AI-generated content.
The Next Web (TNW) · 85d ago
Meta's Oversight Board has issued recommendations to improve the reporting and mitigation of non-consensual sexualized AI deepfakes. The board emphasizes that current processes are insufficient for protecting ordinary users.
Engadget · 88d ago
The surge in AI-generated child abuse material is overwhelming law enforcement agencies. Investigators are struggling with the psychological toll and lack of adequate support systems to handle this new wave of digital evidence.
Bloomberg Technology · 88d ago
A new report from Chayn argues that tech companies and authorities are misplacing their focus by prioritizing nudity detection over the fundamental issue of user consent. The findings highlight the complex reality of online abuse in the age of generative AI.
BBC Technology · 90d ago
A report by Kapwing reveals that nearly 60% of videos served to new TikTok accounts are AI-generated content. The study analyzed over 10,000 videos across 20 categories to assess the prevalence of automated media on the platform.
The Next Web (TNW) · 90d ago
YouTube has intensified its enforcement of inauthentic content policies, leading to the termination of 16 major channels. The move, aimed at curbing AI-generated 'slop,' is inadvertently penalizing human creators who do not appear on camera.
The Next Web (TNW) · 96d ago
YouTube is implementing stricter measures against low-quality AI-generated content, which is inadvertently affecting 'faceless' creators. These creators, who built audiences without appearing on camera, are now facing monetization challenges due to the platform's new enforcement policies.
Digital Trends · 96d ago
The Australian Human Rights Commission is urging for a 'digital duty of care' to address how social media algorithms amplify racist content. This follows reports of harmful, racially offensive videos targeting First Nations people spreading rapidly across platforms.
The Guardian Technology · 99d ago
Deezer has released a new tool that allows users to scan playlists from any streaming platform to identify AI-generated or synthetic music. This move aims to provide transparency in the music industry, as other major platforms have yet to adopt standardized AI labeling.
The Verge · 100d ago
The 'Query Retrieve Conclude' framework addresses the challenge of interpreting dynamic, emerging memes by retrieving real-time web evidence. It outperforms existing zero-shot baselines in both meme understanding and detection tasks.
ArXiv AI · 105d ago
Synthesia has integrated Cinder's infrastructure to scale its moderation capabilities. This allows the platform to screen and moderate AI-generated video content before the rendering process begins.
The Next Web (TNW) · 107d ago
YouTube is implementing an automated detection system to identify and label content created with artificial intelligence. This initiative aims to increase transparency for viewers regarding AI-generated media.
Engadget · 115d ago
YouTube is enhancing its AI disclosure system by increasing label visibility and introducing automated AI content detection. These changes are scheduled to roll out by May 2026 to improve transparency for viewers.
Digital Trends · 115d ago
YouTube is implementing automated labeling for videos containing significant photorealistic AI content. This move reduces reliance on creator disclosure and increases transparency for viewers.
TechCrunch AI · 115d ago
The popular web novel platform 'Shosetsuka ni Naro' will require authors to report AI usage in their works starting June 9. Failure to set these parameters will result in a suspension of posting privileges from September.
ITmedia AI+ (日本) · 115d ago
Universal Music Group and TikTok have renewed their partnership with a primary focus on implementing stricter moderation policies against unauthorized AI-generated music. This agreement aims to protect artist rights and intellectual property in the era of generative AI.
TechCrunch AI · 116d ago
LinkedIn is introducing new systems to identify and limit the reach of generic AI-generated posts and automated comments. This move aims to prioritize authentic human interaction and reduce the prevalence of low-quality AI slop on the platform.
Digital Trends · 122d ago
Spotify is implementing new safeguards to label AI-generated content on its platform. This initiative aims to address growing concerns regarding impersonation, fraud, and the authenticity of music in the streaming era.
TechCabal · 123d ago
Google is facing accusations of violating the UK's Online Safety Act by continuing to display search results for a US-based suicide forum. The site has been linked to 164 deaths in the UK and was previously fined £950,000 by Ofcom for failing to restrict access.
The Guardian Technology · 128d ago
Meta is using AI to scan photos and videos on Facebook and Instagram for bone structure and height to detect users under 13. The system avoids facial recognition, focusing on general visual cues, and also checks contextual clues in posts, comments, bios, and captions.
The Verge · 137d ago
YouTube is expanding its AI likeness detection technology to talent agencies, managers, and celebrities. This tool helps identify and remove deepfakes featuring their likenesses.
Digital Trends · 151d ago
YouTube is expanding its likeness detection technology to the entertainment industry. This AI tool helps celebrities and teams identify and remove AI-generated content and deepfake videos from the platform.
cnBeta (Full RSS) · 151d ago
YouTube is expanding its AI likeness detection technology to celebrities. The tool allows talent and representatives to identify and remove deepfakes.
TechCrunch AI · 151d ago
Deezer receives 75,000 AI-generated songs daily, comprising 44% of all uploads. AI tracks account for just 1-3% of total streams.
The Verge · 152d ago
Nairobi firm Sama abruptly sacked over 1,000 low-paid Kenyan workers after Meta ended their outsourcing deal. Termination followed allegations of staff viewing private smart glasses footage.
The Guardian Technology · 155d ago