WhatsApp Channel Join Now
Telegram Group Join Now

Relevance: GS Paper II — International Relations (technology & strategic autonomy); GS Paper III — Science & Technology Source: News reports & Anthropic statement, June 2026

1 · What happened

The US government ordered the American company Anthropic — maker of the Claude AI assistant — to cut off access to its two most advanced AI models, Fable 5 and Mythos 5, for every foreign national. The order was issued by the US Commerce Department on national-security grounds.

Because the company could not quickly separate foreign users from American users, it had to switch both models off for everyone, worldwide. All other Claude models keep working normally. Anthropic has said it disagrees with the order and is trying to restore access. India is among the most affected, as it is one of the largest users of Claude outside the US.

2 · How the story unfolded

Export controls: rules a government uses to stop sensitive technology from reaching others. They focus on “dual-use” technology — things that can serve both everyday and military purposes.
1
Two very powerful AI models
Anthropic’s Fable 5, and the stronger, non-public Mythos 5 it is built on, are among the most capable AI systems in the world today.

2
A safety concern appears
These models can find weak spots in computer systems very quickly. Someone found a way — a “jailbreak” — to unlock this risky ability in Fable 5 using clever instructions.

3
The US government steps in
Worried that rivals could misuse this power, the government ordered both models blocked for every foreign national. As the company could not tell users apart in time, it shut the models off worldwide.

4
India feels the shock
India is the second-largest user of Claude after the US (5.8% of global use). Indian startups and researchers building on these models lost their foundation overnight.

5
The lesson — build our own
The episode shows the danger of depending on foreign technology. India is pushing for “sovereign AI” — its own models and computing power — through the IndiaAI Mission.

  • From chips to software: Until now, US controls mainly stopped the sale of physical AI chips. Restricting a finished AI model is new, and marks a sharp rise in the technology rivalry between major powers.
  • Anthropic’s view: The company argues the flaw was narrow, that other public models (such as OpenAI’s GPT-5.5) can do the same thing, and that withdrawing a tool used by millions is too strong a step.
  • A warning for developing nations: Poorer countries supply large amounts of data that help train foreign AI. If access to that AI can be cut off at any moment, they risk giving away their data with no guarantee of benefit in return.

UPSC Value Box
Fable 5 / Mythos 5 Anthropic’s most advanced AI models; Mythos is the restricted full model, Fable 5 the public version built on it.
Export controls Government rules limiting the transfer of sensitive technology.
Dual-use technology Technology useful for both civilian and military purposes.
Frontier AI The most advanced AI systems available at any given time.
LLM (Large Language Model) An AI trained on huge amounts of text to understand and produce language.
GPU A powerful chip used to train and run AI systems.
Jailbreak Using clever prompts to push an AI past its safety limits.
Distillation Building a smaller AI by learning from a bigger model’s answers — a way to copy its skill.
Sovereign AI A country’s own AI models, chips, and data, kept under its control.
IndiaAI Mission Approved 2024; ₹10,371.92 crore over 5 years; 10,000+ GPUs; under MeitY (Ministry of Electronics & IT).
Anthropic Economic Index Anthropic’s usage study; India = 5.8% of global Claude use, second only to the US.
DPDP Act, 2023 India’s Digital Personal Data Protection Act; relevant to linking data sharing with technology access.

MCQ Practice Question
Q. With reference to the recent US action on Anthropic’s AI models, consider the following statements:

  1. The IndiaAI Mission, approved in 2024, plans a public computing capacity of more than 10,000 GPUs.
  2. The US order blocked the two AI models only for foreign nationals living outside the United States.
  3. Earlier US export controls mainly covered AI hardware such as advanced chips, while this action targeted an AI model itself.

Which of the statements given above is/are correct?
(a) 1 and 2 only    (b) 2 and 3 only    (c) 1 and 3 only    (d) 1, 2 and 3

Answer: (c) 1 and 3 only

  • Statement 1 — Correct: The IndiaAI Mission (approved 2024) plans a shared computing capacity of more than 10,000 GPUs.
  • Statement 2 — Incorrect (the trap): The order applied to any foreign national, whether inside OR outside the US — even the company’s own foreign-national staff. Limiting it to those “outside the United States” is wrong; that wider scope is why the models had to be shut off worldwide.
  • Statement 3 — Correct: Earlier controls mainly targeted hardware like advanced chips; restricting a finished AI model is a new and far-reaching step.
Ajmal IAS APSC Foundation Course

Recent Posts

Ajmal IAS APSC Foundation BATCH
Ajmal IAS APSC Foundation Course

Start Yours at Ajmal IAS – with Mentorship StrategyDisciplineClarityResults that Drives Success

Your dream deserves this moment — begin it here.