Google, OpenAI, Microsoft and Anthropic all said on 30 October that they plan to support one shared line in robots.txt that tells AI companies not to train on a website. Once each one supports it, a site owner can say no to training at all four with a single rule.
One robots.txt rule to opt out of AI training
Google's post says website owners will be able to use a shared "policy token" in robots.txt to set their training preference across participating AI providers. If a site opts out this way and doesn't mention Google-Extended, Google plans to treat it as if Google-Extended were disallowed, and it aims to support the control in 2027. The token is __usage_ai-training, set out in an Internet-Draft by Anthropic's Kevin Kelley. It isn't yet an endorsed IETF standard. OpenAI aims to support it in the coming months, Microsoft from 1 February 2027, and Anthropic says it has started implementing it, Unite.AI reports.
The draft's example blocks training everywhere except a press-kit folder:
User-Agent: __usage_ai-training
Disallow: /
Allow: /press-kit/
Why it matters: you can keep your pages in search and still say no to training. It's a stated preference, though. A crawler that doesn't support the token will ignore it, and it doesn't pull back content already collected.
A tampered Claude Code extension on Cursor tried to plant SSH backdoors
Cursor says that between about 2:27am and 1:35pm Pacific time on 29 October, versions 2.1.312 and 2.1.313 of the Claude Code extension for Windows and macOS in its marketplace ran code that tries to install a persistent SSH backdoor. Cursor says Anthropic didn't build those versions. Cursor removed them, shut the whole marketplace down, rotated its keys and rebuilt it. The code lacked the admin rights to switch on SSH on most machines, but uninstalling may not be enough. BleepingComputer puts the reach at up to 5,000 developers.
Why it matters: AI coding tools run with the same access as the person using them. If your developers use Cursor, ask them to check today, and keep a list of which extensions are allowed.
AI stocks slide as spending worries grow
Alphabet, Amazon, Meta and Microsoft have laid out plans to spend more than $740 billion on capital spending this year, TechCrunch reports. In US trading on Friday 30 October, the Nasdaq was down 2.9%, Nvidia 4% and Qualcomm 11% when TechCrunch reported, a day after Alphabet's worst day on record. Of the four, only Microsoft still brought in more cash than it spent last quarter. Together with Oracle, they have raised $326 billion in debt in 2026, three times last year's figure, according to S&P Global Market Intelligence. We covered Alphabet's cash squeeze on 29 October; now the selling has spread across the group.
Why it matters: the companies selling you AI are under pressure to show it pays. Keep an eye on price and plan changes for the AI tools you rely on.
Google is bringing ads in AI Mode to India
Google said on 28 October that it is scaling ads in AI Mode to India and nine other countries, with Brazil, Mexico and the UK planned for the fourth quarter. Ads will carry a "Sponsored" label. Google says click-through rates are already similar for sponsored and organic product listings inside AI Mode, based on its own US data.
Why it matters: Indian advertisers can now show up inside Google's AI Mode, so check where your search ads are appearing.
IndiaAI's tender for 10,000 more compute units reaches technical bids
C-DAC was due to open technical bids at 4pm IST on 30 October from companies that want to supply up to 10,000 AI compute units for five years under the IndiaAI Mission, Elets CIO reports. Each cluster must sit in a single data centre in India. More than 38,000 GPUs have already been added to the platform at subsidised rates, yet startups and companies still report waits of months for high-end capacity.
Why it matters: more shared compute inside India makes it easier for Indian firms to train or fine-tune models without renting abroad.
Worth watching: the date OpenAI picks for the training opt-out, and whether Cursor or Anthropic explain how a tampered build got onto Cursor's marketplace under Anthropic's name.
Arthat AI
Working on something similar at your company?
Arthat AI builds these automations for Indian businesses every week. If you want to talk through whether this fits your situation, we'll give you a straight answer — even if the answer is "don't build it."
