Earlier this month, Google publicly griped that “commercially motivated” actors were trying to clone its Gemini AI through agents that queried the chatbot up to 100,000 times to “extract” the underlying model.
The hypocrisy of Google’s accusations was palpable. For years, the search giant has relied on indiscriminately scraping the internet for content to train its AI models, without compensating copyright holders — and racking up lawsuits as a result.
Now Anthropic has entered the fray. Unlike Google, the company behind chatbot Claude was willing to point fingers, accusing Chinese AI firms DeepSeek, Moonshot, and MiniMax of “distilling” its AI model. The company claimed in a new blog post that the accused firms created more than 24,000 fake accounts that queried Claude 16 million times, a “violation of our terms of service and regional access restrictions.”
Distillation is essentially when a small “student” model is trained to replicate the performance of a much larger “teacher” model — a convoluted term essentially denoting the act of copying someone’s homework without express permission.
Even ChatGPT maker OpenAI has accused DeepSeek of distilling its AI models in a statement earlier this month, highlighting a broader backlash to Chinese entities reverse-engineering extremely costly AI software — despite the companies themselves amassing intellectual property and feeding it to their models with little regard for securing the required permissions.
In its blog post, Anthropic cried foul and accused the Chinese companies of “illicitly” extracting “Claude’s capabilities to improve their own models.”
“Distillation is a widely used and legitimate training method,” it wrote. “For example, frontier AI labs routinely distill their own models to create smaller, cheaper versions for their customers.”
“But distillation can also be used for illicit purposes: competitors can use it to acquire powerful capabilities from other labs in a fraction of the time, and at a fraction of the cost, that it would take to develop them independently,” the blog post reads.
Some queries the company identified involved asking Claude to “imagine and articulate the internal reasoning behind a completed response and write it out step by step — effectively generating chain-of-thought training data at scale.”
The allegations come as DeepSeek prepares to release its V4 model, an occasion that could rattle the American AI industry. Perhaps the biggest perpetrator, according to Anthropic, though, was MiniMax, a Shanghai-based company that IPO’d on the Hong Kong Stock Exchange last month. Anthropic says it identified “over 13 million exchanges” from it.
“When we released a new model during MiniMax’s active campaign, they pivoted within 24 hours, redirecting nearly half their traffic to capture capabilities from…
Source link
Disclaimer
We strive to uphold the highest ethical standards in all of our reporting and coverage. We blogs.grocliq.com want to be transparent with our readers about any potential conflicts of interest that may arise in our work. It’s possible that some of the investors we feature may have connections to other businesses, including competitors or companies we write about. However, we want to assure our readers that this will not have any impact on the integrity or impartiality of our reporting. We are committed to delivering accurate, unbiased news and information to our audience, and we will continue to uphold our ethics and principles in all of our work. Thank you for your trust and support.
Website Upgradation is going on for any glitch kindly connect at [email protected]
