A seguirPróximo guia
EU Digital Services Act Rules for Recommender Systems
Sociedade
GUIA DA SOCIEDADE
The Digital Services Act regulates intermediary services and online platforms, while the AI Act regulates defined AI systems and actors.
Generative AI can matter under both when it is integrated into a covered platform or affects how content and recommendations operate.
The Digital Services Act (DSA) creates due-diligence duties for intermediary services, with additional obligations for online platforms and the strictest systemic-risk duties for designated very large online platforms and search engines (VLOPs and VLOSEs). It does not generally ban generative AI or decide whether a piece of content is lawful. Instead, it sets processes and safeguards for services that host, distribute, recommend, or search for content. Other law determines what content is illegal. Generative AI may affect a service’s risks in several ways: users may upload synthetic material, platform tools may generate or transform content, and recommender systems may amplify content regardless of its origin. Under DSA Articles 34 and 35, VLOPs and VLOSEs assess systemic risks arising from the design, functioning, and use of their services and take reasonable, proportionate measures tailored to identified risks. Relevant concerns include dissemination of illegal content, negative effects on fundamental rights, civic discourse, electoral processes, public health, and minors. The AI Act addresses different actors and system categories. Where a general-purpose AI model with systemic risk is integrated into a VLOP or VLOSE, both laws may be relevant, but their risk frameworks are distinct. The AI Act considers model-level risks and provider duties; DSA assessment considers service-level risks connected to reach, design, operation, and use. One cannot assume that satisfying one framework resolves every obligation under the other. For a platform adding an AI image generator or chatbot, the first steps are to determine its DSA service category and whether it has a VLOP/VLOSE designation, then map how the feature changes content distribution and user exposure. The platform should examine relevant risk evidence and consider proportionate mitigations such as interface, moderation, recommender, or internal-process changes. Whether a particular measure is required depends on the identified risks and law. The DSA and AI Act should be read alongside data protection, consumer, copyright, and other applicable rules.
Os danos catastróficos e diários da IA dependem de quem entende os riscos e de quem pode agir.
A literacia pública e profissional determina se uma política de segurança forte é politicamente possível.
Explicações claras reduzem a captura por exageros, relações públicas de laboratório e teatro de ética vaga.
Platforms are likely to improve risk inventories to capture generative features, synthetic-content flows, and recommender effects across languages and communities. Better measurement may connect launches to user exposure and observed harms, helping providers choose proportionate interventions. The DSA requires periodic assessment and review; no single detection tool solves systemic problems. As products change, providers should reassess before material features launch and preserve evidence about mitigation outcomes for later review. Guidance may continue to clarify how services document emerging harms. Teams should record why selected measures fit the evidence.
A designated very large platform assesses whether a new AI image feature could change the spread of harmful content and updates its service-level risk assessment.
A social network checks how recommender ranking amplifies synthetic election claims and evaluates proportionate changes.
A chatbot provider and a platform separately map their AI Act provider duties and the platform’s DSA obligations.
A marketplace records that the DSA process governs service due diligence while product-safety rules may separately govern listed goods.
Tratar o risco existencial como ficção científica enquanto aumenta a capacidade.
Confundir segurança do produto de superfície com alinhamento sob alta autonomia.
Deixando o público não-inglês e não especializado com apenas fontes de baixa qualidade.
Separe os riscos de danos ao produto, uso indevido e perda de controle/desalinhamento.
Pergunte quais evidências mudariam sua visão sobre prazos e gravidade.
Prefira fontes primárias e avaliações concretas em vez de afirmações de marketing.
Identifique um caminho de ação: carreira, política, financiamento ou habilidades – não apenas conscientização.
Free newsletter
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
The Digital Services Act regulates intermediary services and online platforms, while the AI Act regulates defined AI systems and actors. Generative AI can matter under both when it is integrated into a covered platform or affects how content and recommendations operate.
The DSA addresses online intermediary services and their obligations.
These systemic-risk duties apply to designated very large platforms and search engines.
The DSA assessment concerns the online service and its systems.
AI can change risks, but does not determine platform status or legality by itself.
Article 34 describes recurring assessments and pre-launch assessment for critical functionalities.
Continue aprendendo
Mais guias escolhidos para este tópico
A seguirPróximo guia
EU Digital Services Act Rules for Recommender Systems
Sociedade