Anthropic says new protections have significantly reduced the amount of data stolen by "distillation attacks" Anthropic said on Monday that new measures aimed at preventing competitors from copying the capabilities of its Claude AI model have reduced the amount of data stolen by so-called "distillation attacks" by about 90% to 99%.

Anthropic says new protections have significantly reduced the amount of data stolen by "distillation attacks" Anthropic said on Monday that new measures aimed at preventing competitors from copying the capabilities of its Claude AI model have reduced the amount of data stolen by so-called "distillation attacks" by about 90% to 99%. Model distillation has become a focus for top AI companies; companies such as Anthropic, OpenAI and Google believe the practice enables competitors to replicate their advanced capabilities faster and at lower cost. Anthropic said that protective measures aimed at restricting access to Claude’s inference capabilities have significantly reduced extraction attempts against the model, and that the “number of inference tokens extracted through distillation” has now dropped to “1/100 to 1/10 of the previous level.” The California-based company said it has made it more difficult to exploit the proxy network to repeatedly probe the Claude protected inference process, blocking data extraction attempts and hampering large-scale model distillation campaigns. Anthropic said its security system is "roughly equivalent to a system that would prevent Claude from assisting in the creation of a biological weapon" and has a "false positive rate of less than 0.1 percent." via cnBeta.COM - Chinese industry information station (author: Source: Global Market Report)