AI News Digest

Archives
Log in
Subscribe
August 5, 2026

Meta Removed Over 50 Paid Ads Featuring AI-Generated Sexual Images of Children

1. Meta下架逾50条含AI生成儿童性虐待图像的付费广告,部分曾触达数千账户 Meta removed more than 50 paid advertisements containing AI-generated child sexual abuse material or sexualized images of minors after WIRED contacted the company about findings from the Tech

2. 国际刑警组织称非洲36国记录的网络犯罪案件中55%涉及AI Artificial intelligence was involved in 55% of the cybercrime cases recorded by 36 African countries covered in INTERPOL’s African Cyberthreat Assessment Report 2026.

3. Hassabis卸下Google DeepMind日常管理,Kavukcuoglu接掌Gemini与前沿研究 Google is handing day-to-day control of Google DeepMind, its central AI research and development organization, to longtime executive Koray Kavukcuoglu.


In Brief

  • Anthropic forms custom AI chip design team Claude maker Anthropic confirmed it is hiring engineers for a custom silicon team, aiming to co-design models and hardware for faster, more efficient AI workloads.
  • Meta launches Muse Code for large repositories Meta released Muse Code, a beta terminal agent powered by its Muse Spark coding model that can plan, implement, and validate changes across large codebases. It can divide larger jobs among parallel sub-agents working in isolated worktrees.
  • Google schedules mobile Assistant shutdown for September Google reportedly told users it will begin removing Google Assistant from Android phones, tablets, and paired devices on September 4, leaving Gemini as the available assistant on supported devices and in eligible regions. Google Home and Google TV devices are not currently included.
  • Reddit expands LLM-powered moderation to new communities Reddit’s optional Rules Hub uses large language models to evaluate whether posts and comments violate the intent of community rules. After tests in more than 700 communities, Reddit is extending it to all newly created communities ahead of wider availability later in 2026.
  • Shopify reports AI-referred traffic and orders tripled Shopify said traffic and orders attributed to AI tripled year over year in its second quarter, while traditional search still generated roughly one-third of storefront sessions. The company also said half of AI-referred visits landed directly on product pages.
  • Hark opens waitlist for browser task agent Hark, an AI startup, previewed Hark Handoff, a browser-use agent designed to complete tasks on websites without official APIs by interpreting page structure and visual information. The company plans to release it by the end of summer, but its demonstration showed only part of a task.
  • Spotify adds independent labels to consent-based AI remix project Spotify partnered with music-licensing organization Merlin, bringing a network of more than 30,000 independent labels and distributors into its planned AI covers and remixes product. Participating artists must consent and are intended to receive credit and compensation; Spotify has not announced the preview date.
  • US frontier-model review framework excludes open models According to Axios, the Trump administration’s voluntary framework gives the government 30 days to review certain unreleased closed models but excludes open models. The reported framework does not define “state-of-the-art” capabilities or “national security risk,” and its details are not planned for public release.
  • Klaviyo acquires AI customer-success startup Agency Marketing automation company Klaviyo agreed to acquire Agency, an AI-powered customer-success startup, for undisclosed terms. Agency founder Elias Torres will become Klaviyo’s chief product officer, and its 25-person team will help develop Klaviyo’s campaign-building and post-sale support agents.
  • MacPaw and Liquid AI develop local inference stack MacPaw partnered with model developer Liquid AI to build Elix, an on-device inference and memory system for MacPaw’s Eney assistant. MacPaw plans eventually to offer the stack to developers building apps for its Setapp subscription store.
  • JoyAI system reports real-time 720p video editing Researchers introduced JoyAI-Video-Edit, a 16-billion-parameter autoregressive diffusion system for editing ongoing video without knowing future frames or the final duration. They report approximately 30 frames per second at 720p on one Nvidia B200 GPU and have released the code.

Read the full edition →

Don't miss what's next. Subscribe to AI News Digest:
← Newer DeepMind Opens WeatherNext Models, Claims More Than a Day’s Cyclone Forecasting Gain Older → Officials Order 58,000 Exam Retakes After Top Scores Jump Fivefold
Powered by Buttondown, the easiest way to start and grow your newsletter.