The tug of war between China and the USA in the field of artificial intelligence continues. In the race for technological supremacy between the two countries, important Chinese artificial intelligence laboratories are accused of “systematic” theft of American technologies.
Read more Brexit, EU refunds arrive: 44 million for the Customs Agency
Ask the Sun
The questions are automatically suggested by 24Ore AI
based on the content viewed.
The National Security Agency (NSA), the Cybersecurity and Infrastructure Security Agency (CISA), and the Federal Bureau of Investigation (FBI) have indeed drafted a cybersecurity warning document, investigating the activities of some Chinese companies regarding the systematic theft of functionalities and capabilities of American models.
The accusation: copied functionalities and distillation techniques
These practices – the document explains – occur “through industrial-scale knowledge distillation campaigns that constitute the core – and not a mere complement – of their (Chinese industries, ed.) AI development strategy.”
Read more Energy transition: Lombardy focuses on sustainability, innovation, and infrastructure
Distillation, in itself, is a widely used machine learning technique in the sector that allows transferring knowledge from a large pre-trained model, the “teacher model,” to a smaller “student model.” However, Chinese companies would be making “malicious” and “aggressive” use of it, benefiting from significantly shorter AI development times and reduced financial costs for training cutting-edge models, with the goal of “bridging the technological and performance gap” between the computing assets of the two countries.
The companies involved
The joint US investigation names: DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI as the Chinese companies responsible for the systematic theft.
Read more Advertising: Netflix, Amazon, Disney, and Google in Audicom for measurements
