Automated trust-and-safety at scale: classifiers, deepfake detection and policy enforcement.
266 articles
WIRED examines how AI-generated sexualized deepfakes are targeting teachers in schools, extending the harm beyond students. The affected educators describe difficulties identifying perpetrators and securing meaningful accountability.
Wired AI · 26d ago
LinkedIn’s “Seems like AI slop” reporting button has been clicked by more than one million users since its July 30 launch. The feature lets users report suspected low-quality AI-generated content through the three-dot menu on a post.
cnBeta (Full RSS) · 29d ago
TechCrunch testing found that Anthropic’s Claude models, including Opus 4.6, could be prompted to generate sexually explicit content despite the company’s prohibition. The findings suggest that the model’s safety restrictions may be vulnerable to relatively simple prompt-based circumvention.
TechCrunch AI · 29d ago
LinkedIn says more than one million users have clicked its “Seems like AI slop” reporting button since its July 30 launch. The feature accompanies improved AI-content classifiers and follows findings that 41% of LinkedIn long-form posts were flagged as fully AI-generated.
The Verge · 29d ago
Apple Music appears to be focusing more closely on AI-generated music and does not want such tracks to escape its oversight. The update could affect how AI-created songs are identified, labeled, or distributed on music platforms.
Ifanr (爱范儿) · 29d ago
Wired reported that Meta’s Facebook and Instagram displayed ads for an AI pornography-generation tool. One ad featured a deepfake pornographic video resembling a prominent American female politician, and Meta removed the ads after being questioned.
cnBeta (Full RSS) · 30d ago
Meta reportedly ran advertisements for an app that promised to digitally undress female politicians. One ad included a pornographic video featuring a deepfake closely resembling a US politician.
Ars Technica AI · 31d ago
The article warns that AI-generated faces, cabins, uniforms, and lip-synced videos can impersonate airline staff and make flirtation-based scams more convincing. It urges airlines and platforms to act against unauthorized use of uniforms, logos, and employee identities before the scams expand beyond small, repeated payments.
虎嗅 · 32d ago
OpenAI is launching ChatGPT for users aged 13 to 17, with stronger safeguards around suicide, self-harm, and romantic or sexual conversations. The product also includes homework and study support intended to help students learn instead of simply generating finished answers.
The Guardian Technology · 32d ago
OpenAI announced a ChatGPT mode designed for teenagers. The mode will automatically limit certain conversations to provide additional protection for young users amid growing safety concerns.
New York Times Technology · 32d ago
An analysis from the Global Network on Extremism and Technology has been cited as evidence that violent AI-generated content causes cognitive decline and violence in children. TNW says the underlying report is more cautious and uncertain than the viral headlines suggest.
The Next Web (TNW) · 32d ago
The New York Times reported that low-quality AI-generated content is spreading across Google Search, Amazon product pages, and Facebook feeds. Examples include absurd recipes, fabricated biographies, and bizarre AI-generated images, prompting technology companies to confront the problem.
cnBeta (Full RSS) · 33d ago
Rogue Studios is a new cinematic adult AI generator focused on producing sophisticated adult content. The company is positioning itself as a premium, HBO-like studio rather than a general-purpose generation tool.
Wired · 33d ago
Spotify, LinkedIn and other platforms are taking steps to address the growing volume of low-quality, AI-generated content. The trend highlights increasing pressure on digital services to improve content quality and user trust.
New York Times Technology · 33d ago
China Poetry Net has integrated DeepSeek into its editorial workflow, automatically generating commentary for every uploaded poem and replacing traditional expert reviews in its Daily Good Poem column. Editors still review, shorten, and refine the outputs, while acknowledging recurring praise, formulaic interpretation, and ethical concerns around AI-generated literary work.
虎嗅 · 33d ago
Instagram promised to take stronger action against harassment recorded with Meta’s AI glasses. A month later, many of those videos reportedly remain visible on the platform.
Digital Trends · 34d ago
A woman claims her stepfather used Grok to transform a childhood photograph into explicit imagery. She warned that AI tools can turn ordinary personal images into child sexual abuse material.
TechCrunch AI · 35d ago
A 17-year-old Massachusetts teenager has been charged with murdering his mother and younger brother. Prosecutors allege that he used ChatGPT beforehand to explore fantasy scenarios involving the killing of his family, though the excerpt does not establish that ChatGPT caused or enabled the violence.
The Next Web (TNW) · 36d ago
Meta and TikTok agreed to fact-check content related to Spain’s border crossings through a voluntary commitment brokered by Brussels. The move follows the Ceuta tragedy, highlighting how viral videos can rapidly influence real-world events.
The Next Web (TNW) · 39d ago
Spotify will begin labeling artist profiles that appear to represent non-human AI identities with an “AI Persona” badge. Starting in mid-September, music from these profiles will be removed from user recommendations, with detection supported by human review and AI tools.
The Verge · 39d ago
Spotify forecast third-quarter operating income of €670 million, below Wall Street’s €678 million estimate. The company is also dealing with slower user growth and a surge of AI-generated tracks across streaming services.
The Next Web (TNW) · 46d ago
X has reached a quiet settlement with a group of music publishers regarding a 2023 lawsuit. The original complaint alleged that the platform hosted thousands of instances of copyright-infringing content.
Engadget · 63d ago
The article analyzes how Brazilian funk music, containing explicit lyrics, bypasses Chinese content moderation systems because the language is not understood by algorithms or local users. This highlights a significant blind spot in AI-driven content safety where semantic understanding fails across different linguistic contexts.
虎嗅 · 66d ago
A judge has dismissed a lawsuit that accused Apple of failing to adequately limit the storage and distribution of child sex abuse material on its iCloud service. The case centered on the tension between user privacy and the platform's responsibility to monitor illegal content.
New York Times Technology · 66d ago
A US federal judge dismissed a class-action lawsuit against Apple regarding child abuse imagery in iCloud. The court ruled that Apple is protected by Section 230 of the Communications Decency Act.
The Next Web (TNW) · 66d ago
UK regulator Ofcom has issued significant fines to multiple adult content platforms for failing to implement effective age verification systems. This move signals a stricter regulatory environment for platforms handling sensitive user data and age-restricted content.
BBC Technology · 72d ago
Scammers are hijacking government domains to host malicious OnlyFans-related ads. Ironically, mass copyright takedown requests from adult creators are forcing these sites offline, inadvertently cleaning up the web.
Wired · 73d ago
The Indian Ministry of Information and Broadcasting has ordered Telegram to take stricter action against piracy channels. The platform must submit an action-taken report within 15 days.
cnBeta (Full RSS) · 76d ago
Fanfiction communities are actively attempting to identify and purge works created with generative AI tools like ChatGPT and Claude. The movement faces criticism due to the unreliability of current detection methods, which may unfairly target human authors.
The Verge · 77d ago
Spain's football league, LALIGA, has inadvertently blocked over 500,000 legitimate websites while attempting to stop illegal streaming. The incident raises concerns about the collateral damage caused by automated content-blocking systems.
The Next Web (TNW) · 78d ago