Tag
The AI short drama "Armed Entry into Song" achieved over 6 billion views with a small team of about ten people using a "one-person production" model, leveraging Seedance 2.5, AI dubbing, and JianYing for production, demonstrating a viable path toward industrialized AI content creation.
Google has launched its experimental AI app Dreambeans in India, its first market outside the U.S., with support for 13 local languages, aiming to bring personalized AI-generated visual stories to a broader audience.
A Reddit user asks how people spot AI-generated posts and comments, with commenters noting that recognizing the stylistic quirks of AI writing is the main way to tell.
AI-generated fan wikis full of errors are crowding out legitimate game wikis in search results, frustrating players and forcing publishers like Hooded Horse to fight a 'whack-a-mole' against mass-produced low-quality sites.
Matt Pocock complains that AI-generated replies on the platform have become so bad that he no longer checks notifications, instead filtering for profile pictures of people he knows aren't bots as they get harder to spot.
A user comments that they've never seen so many AI-generated replies on X, calling it the worst it has ever been, reflecting growing concern over AI bot activity on the platform.
The US Department of Education's Harry Potter-style X post appeared to promote the 'Malfoid,' a right-wing fan-fiction meme featuring a gender-flipped, blood-purity-obsessed female Draco Malfoy that has become a vehicle for far-right aesthetics and AI-generated propaganda.
Fenng and others criticize the economics of AI-generated drama series: one AI drama racked up over 4 billion cumulative views, with compute costs exceeding a million yuan yielding only tens of millions in returns, prompting doubts over whether the business model for churning out such low-quality content is worthwhile — with some even joking that writing web novels offers better returns.
UC Berkeley 和华盛顿大学的一项新研究发现，到 2026 年 8 月，约 24% 的可抓取网页包含 AI 生成或 AI 编辑的文本，相比 2022 年前几乎为零大幅增长；小而低链接的网站贡献了 78.8% 的增长。
Martin Fowler discusses the phenomenon of "LLM Voice" in prose — the uncanny-valley feeling readers get from AI-written text — and urges authors to avoid letting LLMs draft their writing, warning that delegating writing also delegates thinking.
This arXiv paper analyzes over 5 million webpages from 2019 to 2026, finding that the share of pages with AI-generated text grew from near zero before 2022 to nearly 25% by August 2026. The rise is uneven: English-language and .com domains have higher AI prevalence than other languages or .gov/.edu domains, and AI text concentrates in the least central pages of the web graph, with estimates produced via a Bayesian hierarchical model correcting for detector error rates.
A Guardian review harshly criticizes Elon Musk's AI-generated seven-minute Odyssey episode made with Grok Imagine, calling it passionless, visually inert, and a pointless riposte to Christopher Nolan's successful film adaptation.
China's first fully AI-generated theatrical feature, Bona Film Group's 'Sanxingdui: Future Past', flopped at the box office with just $181,000 in earnings, two months after the scrappy low-budget animated film 'Niu Lai' became a surprise $10 million hit.
Venus-DefakerOne2 is a unified anti-forgery vision foundation model that broadens detection across forgery forms (digital manipulation, physical spoofs) and media types (images, videos) using Tide-Style-Tuning and entropy-based test adaptation, achieving state-of-the-art results on multiple forensic benchmarks.
法国总统马克龙批评世界领导人（暗指特朗普和内塔尼亚胡）基于AI生成的虚假视频对法国骚乱发表“愚蠢”言论，并指出存在外国信息操纵。
An informal experiment testing multiple AI detectors found they produce inconsistent and unreliable results, flagging human-written essays as AI while sometimes missing actual AI-generated text. The author argues this unreliability is a serious problem for students and submitters whose work may be judged by these tools.
A tweet criticizing CNN's Erin Burnett for complaining about the flood of AI-generated translated content on X, reframing it as a business crisis for CNN rather than for X.
法国政府就骚乱紧急召集TikTok、eBay、Amazon、Meta等六大平台开会，要求加强内容审核并打击赃物转售，同时应对AI生成的抗议视频传播。
研究者Peter Henderson在ICLR审稿中表达对AI冲击学术同行评审体系的担忧，指出大量疑似AI生成的投稿正导致审稿质量下降与评审系统的死亡螺旋。
Thousands of AI-generated comments misidentified Albania as France and flooded comment sections with near-identical posts blaming Muslim immigrants for protests and advocating lethal police force, highlighting growing issues with AI-driven misinformation.