Abstract: The intelligent algorithm recommendation model for personalized users has been widely used in major network service platforms, which effectively solves the problem of information overload in the era of nationwide creation, but also leads to difficulties and disputes in determining copyright contributory infringement. The platforms tend to be “pseudo-neutral” and value-oriented from the perspective of their operation mechanism and their practice of intelligent algorithm recommendation, so it is difficult to be exempted from liability through technology neutrality, which constitutes “active recommendation” in the determination of “should-know” of copyright contributory infringement. Based on the functions of content identification, filtering and recommendation of the intelligent recommendation algorithm, as well as their strong information management ability, the platforms should undertake certain obligations of cencorship and filtering according to its types of service, behavior and objects; otherwise it will be liable for infringement if it violates the duty of care and constitutes “should know”.