Recent news and analysis on Artificial Intelligence

Breaking AI news, ChatGPT developments, machine learning breakthroughs, LLM advancements, and expert analysis on how AI is transforming technology and society.

• Google's Gemini app surpasses 1 billion monthly users

  Google's Gemini chatbot surpassed 1 billion monthly users, making it Google's fastest-growing product ever, CEO Sundar Pichai announced on X. The milestone was reached last week, according to a Google spokesperson. OpenAI's ChatGPT also recently passed 1 billion monthly users, as announced in an August 6 blog post. The achievement comes amid a major reorganization of Google's AI division, with key researchers leaving, and while rivals have launched more powerful models.

• Spotify to label AI-generated songs with 'AI Personas'

  Spotify will launch an 'AI persona' badge in mid-September to identify AI-generated content producers, addressing concerns over AI music overwhelming streaming catalogues. Artists can self-disclose starting August 11, otherwise Spotify will review and label profiles. The policy applies to artists whose music is primarily generated by artificial intelligence. The move follows its April 'verified by Spotify' badge and aims to ensure listeners know the origin of the music.

• Anthropic to watermark all Claude AI content

  Anthropic will embed machine-readable watermarks in text, text files, and images generated by Claude, plus cryptographically signed provenance metadata on certain file types, to comply with the EU AI Act's transparency rules. The marks apply globally across Claude platforms, including via AWS, Google Cloud, and Microsoft Foundry. Users worry Claude-edited human work might be falsely flagged as AI-generated, though Anthropic notes a detected mark confirms processing but not full provenance.

• SpaceXAI releases Grok 4.6, matching GPT-5.6 Sol

  SpaceXAI launched Grok 4.6, which Elon Musk claims is "objectively #1" in AI considering intelligence, speed, and cost. The model matches frontier benchmarks at 60%+ lower cost than Claude Opus 5, with pricing at $2/$6 per 1M tokens versus $5/$25 for Claude and $5/$30 for GPT-5.6 Sol. This threatens inference margins for Anthropic and OpenAI.

• China-linked hackers launch unprecedented AI cyber attack on Taiwan

  Suspected China-linked hackers used open-source AI tools in the first observed end-to-end autonomous cyberattack against a government target, compromising at least 85 user accounts and stealing over 2,500 personnel records from Taiwanese systems, according to Israeli cybersecurity firm Dream. The attack expanded to Taiwan's nuclear safety agency and energy companies.

• SpaceXAI unveils Grok Bot, AI agent team for professional tasks

  SpaceXAI launched an early beta of Grok Bot, a persistent AI agent that operates in a cloud-based environment to complete multi-step workplace tasks. Priced at $120 per seat per month for teams and $200 for individuals, it can sign into apps and be messaged like a colleague. Multiple bots run in parallel and coordinate via group chat.

• AI firms shift to simulated workplaces for agent training

  AI companies are moving beyond chatbots to train agents in realistic digital workplaces. Google is in talks to invest over $1.5 billion in Mechanize, while Meta collects employee keystrokes and mouse movements. Scale AI reports nearly half of new projects use reinforcement learning environments.

• AI-generated series 'Castle Walls' premieres on Prime Video

  The AI-generated historical drama 'Castle Walls' launched Wednesday on Prime Video, marking a streaming first as the first show made end-to-end using artificial intelligence. The three-episode series, produced by Ay Yapim's AI Yapim subsidiary, focuses on Iran's Alamut Castle and was built with proprietary AI tools.

• CoreWeave shares jump 17% on record revenue, $104B backlog

  CoreWeave shares jumped 17% premarket after reporting its fifth consecutive quarter of record revenue and a $104.2 billion sales backlog. Revenue hit $2.58 billion for the quarter ended in June, more than doubling year-over-year and beating analyst expectations. The company forecasts third-quarter sales of $3.45 billion to $3.6 billion and secured over $25 billion in new customer commitments after the period ended.

• Twitch adds opt-out toggle for Amazon AI training

  Twitch introduced an opt-out toggle letting streamers prevent Amazon from training generative AI models on their content. The setting is enabled by default, but streamers can disable it under 'Training for Generative AI' in settings. Opting out stops streams, clips, and chats from being used for genAI training, though other AI features may still apply.