2 weeks ago
Weekly AI Wrap: Claude Watermarks, Grok Bot, Gemini Billion Users
This week, a lot of big companies that make smart computer programs called artificial intelligence, or AI, did many new things.
One company called Anthropic started putting secret invisible marks on text written by its AI, Claude.
The marks help people know if a computer wrote the words, because a new rule in Europe says they must.
Another company, xAI, built new helper robots called Grok Bots that can work in the cloud even when your computer is off.
Google said its AI helper, Gemini, now has more than one billion users, and it also showed a new phone called Pixel 11.
In India, leader Narendra Modi promised to teach ten million young people about AI in one year.
China stopped a big company called Meta from buying a smaller AI company named Manus.
Other companies made new AI models that find computer bugs and reason better.
Some French publishers complained that Google's AI summaries took visitors away from their websites.
Music and video services also added rules about AI-made content.
It was a very busy week for AI!
Anthropic added invisible watermarks to Claude-generated text to comply with EU AI Act transparency mandates.
xAI launched Grok Bot in early beta, giving each agent its own cloud-based virtual machine for autonomous work.
Google said Gemini crossed 1 billion monthly active users and unveiled the Pixel 11 lineup with the Tensor G6 chip.
India's PM Narendra Modi announced a program to train 10 million youth in AI over the next 12 months.
China's NDRC blocked Meta's $2 billion acquisition of Manus, which will now operate independently.
- Who
- Anthropic, xAI, Google, OpenAI, DeepSeek, Meta, and Indian Prime Minister Narendra Modi, among others.
- What
- A weekly roundup of AI developments, including watermarking for AI text, new agentic tools and phones, a blocked acquisition, and new AI policies.
- Where
- Global, involving the EU, India, Japan, China, Singapore, France, and the United States.
- When
- The week leading up to August 16, with Modi's announcement made during his August 15 Independence Day address.
- Why
- Motivated by EU transparency regulations, competitive pressure among tech giants, geopolitical oversight, and national AI adoption goals.
Publishers and Regulators
Tech Companies
Google AI Overviews and referral traffic
Publishers and Regulators
Around 300 French publications in APIG filed a complaint with the French Competition Authority, citing an estimated 33 per cent decline in referral traffic and alleging unauthorised content use without collective compensation.
Tech Companies
Google continues to offer Gemini-powered AI Overviews as part of its search experience; no response from Google is reported in the articles.
Meta's acquisition of Manus
Publishers and Regulators
China's National Development and Reform Commission intervened under geopolitical and regional regulatory oversight to block Meta's $2 billion takeover of Manus.
Tech Companies
Meta planned the $2 billion acquisition, while Manus announced it will revert to operating independently and will delete user data integrated since late 2025.
Key facts
- Gemini monthly active users
- 1 billion
- Anthropic watermarking
- Invisible, model-level, preserves readability and resists light editing
- Grok Bot release status
- Early beta, with each agent provisioned its own cloud-based virtual machine
- India AI training program
- 10 million youth to be upskilled in AI over the next 12 months
- Meta-Manus deal
- $2 billion acquisition blocked by China's NDRC
- DeepSeek V4 Pro
- Scored 53 on Artificial Analysis's Intelligence Index (up from 40 for V4 Flash)
- Enterprise-wide gen AI adoption
- Japan 16%, China 98.1%, Germany 91.6%, United States 90.9%
- APIG complaint vs Google
- Around 300 publications; alleged 33% decline in referral traffic and unauthorised content use










