Providers of very large online platforms and very large online search engines should conduct and document risk assessments on the systemic risks stemming from the functioning and use made of their services, in particular as regards risks of illegal content, manipulation and the protection of fundamental rights and of the rights of the child. Those risk assessments should be carried out prior to the deployment of new or significantly modified functionalities that may have an impact on such risks and should be carried out continuously based on the available evidence. Providers should identify the systems they use, in particular algorithmic systems that are used for content moderation and recommender systems, and analyse any potential impact of such systems on the systemic risks identified. Those risk assessments should take into account the different needs of different groups of recipients of the service, in particular minors, persons with disabilities, the elderly and persons at particular risk of being subject to illegal content or content promoting discrimination.