Open-Source AI

Anthropic’s decision to disable access to Fable 5 and Mythos 5 has intensified debate over a far bigger issue than one company’s product policy. For India, the episode has become a warning that dependence on foreign AI platforms can quickly turn into a strategic vulnerability.Zoho founder Sridhar Vembu framed the move as a reminder that technology is now tightly linked to national security and sovereignty. In a post on X, he argued that globalisation has ended and that “Bharat must find her own way ahead”.Why Anthropic shut accessAnthropic said the two models were disabled after it received an order tied to U.S. export controls. Under that direction, the company had to stop access to the models for all foreign nationals, whether they were inside or outside the United States.The restriction also covered Anthropic employees who are foreign nationals. The company said it received the instruction on 12 June and then disabled the models to comply with the requirement.Anthropic also made clear that the action did not affect its entire product lineup. Access to the company’s other AI models, it said, remained unchanged.The shutdown quickly pushed wider conversations in the tech community about export controls, cross-border access to advanced AI, and the risk of relying on a small number of large U.S. providers.Security concerns remain contestedAnthropic said the government did not provide full details about the national security concerns behind the order. But the company said its understanding was that authorities believed there was a way to bypass or jailbreak Fable 5’s protections.Anthropic did not fully agree with that assessment. It said it reviewed the reported techniques and found only a small number of vulnerabilities that were already known.The company added that similar findings can be seen in other publicly available AI models. It also said extensive testing with government agencies, third-party organizations, and internal teams showed Fable 5’s protections were stronger than those of earlier models.Even so, Anthropic said it would comply with the directive while also describing the action as a misunderstanding. The company said it was working to restore access to Fable 5 and Mythos 5 as soon as possible.Vembu pushes a more independent AI pathAgainst that backdrop, Vembu urged India to take a more practical route. He said organizations in India should make greater use of smaller open-source models, including open-source options from India and China.In his view, those models can be made to work effectively with enough effort. He also questioned the long-term value of depending on foreign AI companies when access can be limited at any time by political and security decisions.The episode reflects a broader shift in global AI power. Access to cutting-edge systems is increasingly shaped by geopolitics and national security concerns, rather than market demand alone.Vembu’s criticism was blunt. In another post, he wrote, “Why pay money to people who don’t even want to sell to you?”The cost of building frontier AIVembu also acknowledged that building frontier AI models is difficult. He pointed to the enormous funding required and the challenge of securing advanced GPUs, which are themselves subject to export restrictions.According to Vembu, competing seriously at the frontier would require more than $100 billion in funding. He said that scale is unrealistic and that he does not want governments spending tens of billions of dollars on the effort when the money could be used for more urgent needs.As an alternative, he said Zoho is exploring a much cheaper approach to AI research. He admitted that major breakthroughs take time, but stressed the need for efficient innovations that can be put to practical use.In a separate response on X, he repeated that the immediate task is to make smaller models genuinely useful through real work, not aspiration alone.A broader sovereignty debateThe Fable 5 and Mythos 5 case has strengthened a global debate over who controls the most advanced AI systems. Many countries are now confronting the reality that access to critical technology can be limited unilaterally once national security enters policy decisions.For India, the episode underscores the importance of domestic research, open-source innovation, and locally deployable AI systems. The larger question is no longer only who has the most advanced model, but who can access it, control it, and set the rules for its use. Source: sundayguardianlive.com

Anthropic’s AI Blockage Exposes a Bigger Risk for India’s Tech Future