Changes from report parliamentary committee draft to plenary report

AddedGender bias and discrimination in AI systems

AddedGender-biased training datasets

Added27. Notes that AI training datasets often reflect entrenched stereotypes, resulting in biased recommendation algorithms and content-ordering systems, including the promotion of misogynistic communities, sexualised AI-generated images, pro-anorexia and self-harm content, as well as discriminatory moderation practices; calls for the systematic auditing of training data and algorithmic outputs using gender-sensitive indicators, as well as for scenario testing to examine how recommender systems expose minors to such content, and for platforms to demonstrate measures that prevent harmful content escalation;

Added28. Calls for improved transparency regarding the training of AI models, including clear information on the datasets and materials used, particularly for AI systems used or likely to be accessed by children and young people; stresses that such transparency should be accompanied by child safeguarding measures to prevent exposure to harmful, biased or inappropriate content; calls for the establishment of a code of conduct setting industry standards for transparency, accountability and child protection in the development of these technologies; recalls that the AI Act requires providers of high-risk AI systems to ensure that their training, validation and testing datasets are relevant, sufficiently representative, and to the best extent possible, free of errors and complete, in view of the intended purpose of the high-risk AI systems;

AddedInclusion: biased algorithms and discrimination

Added29. Highlights the need to prevent algorithmic bias and discrimination affecting young people; calls on online platforms to commission regular independent audits and implement transparent reporting mechanisms to monitor and correct systemic bias in AI-driven educational, employment and social media systems;

AddedObligations, transparency and possible solutions

AddedSafeguard measures

Added30. Calls on the Commission, in coordination with the Member States, to develop guidelines on the use of AI systems impacting children, in order to prioritise precautionary measures against engagement-driven design, including protective default settings and understandable explanations of how automated systems work, while ensuring that AI used by children remains safe and available for use;

Added31. Voices concern about AI assistants regularly misrepresenting news content in their outputs; calls on providers of AI assistants to collaborate with the media sector to improve AI assistants’ responses and outputs; calls on the Commission, the Member States and relevant regulatory authorities to effectively apply relevant EU media and digital laws with a view to promoting public interest objectives in AI, such as media pluralism, and to prevent AI-assisted searches from undermining the general public’s trust in news and its access to reliable information;

AddedScams and fraud: specific measures against ‘celebrity scams’

Added32. Expresses concern over the proliferation of AI-generated content impersonating media professionals and brands, for commercial gain or for the dissemination of disinformation or ‘celebrity scams’; urges online platforms to increase mitigation, in this regard, and urges the Commission to classify celebrity scams as a systematic risk and intensify ongoing investigations under the DSA; encourages the Commission to consider adopting more efficient measures, including legislation, to protect minors from deception, while also ensuring effective redress for media organisations and their employees affected by these unlawful practices;

Added33. Expresses concern over dangerous online challenges and disinformation targeting minors; calls for an EU-level rapid alert mechanism in this regard, in cooperation with Safer Internet Centres; calls for better enforcement, based on the current risk assessments by very large online platforms; demands more robust and data-driven systemic assessments to accurately identify and mitigate specific threats to children’s safety;

AddedAI-generated content and deepfakes

Added34. Expresses concern about the growing spread of AI-generated sexualised images, manipulated content and deepfakes designed to deceive, bully, body-shame and manipulate young people, thereby eroding their trust and safety in digital environments;

Added35. Calls for the explicit ban of AI systems generating, manipulating or altering realistic images or videos depicting sexually explicit activities or intimate parts of an identifiable natural person without that person’s consent, and synthetic child sexual abuse material (CSAM), since they are listed explicitly in the prohibited practices under the AI Act; notes that such systems disproportionately target girls and young women, including those in the public sphere, and calls for their deployment, placement on the market or use to be aligned with an ‘unacceptable risk’ classification, so as to ensure consistent enforcement across the EU;

Added36. Calls for the development of comprehensive, accessible and child-appropriate reporting mechanisms and support services that are proportionate, compliant with children’s rights and fundamentally centred on education, prevention and accountability, specifically for minors affected by image-based abuse;

Added37. Calls for specific actions to be taken to prevent intimate partner violence, among and affecting young people, in the digital environment, encompassing acts such as revenge porn, deepfake pornography, hacking, impersonation, and the use of technology to control or limit a partner’s online behaviour; calls on the Member States to better educate young people about this threat;

Added38. Stresses, furthermore, that legislation on gender-based violence must explicitly apply to immersive technologies, such as virtual reality, online gaming platforms and the metaverse, to address non-consensual sexual acts and coerced interactions in virtual spaces;

Added39. Recalls that tackling online sexual abuse and exploitation is essential to addressing the impact of social media on young people; stresses the importance of reinforcing the legal framework, in this regard, and recognises the recast of Directive 2011/93/EUas a key measure for strengthening child protection online;

Added40. Reiterates its position of 16 November 2023 on the proposal for a regulation of the European Parliament and of the Council laying down rules to prevent and combat child sexual abuse (COM(2022)0209), which underlines that detection measures put in place should be targeted, proportionate, subject to judicial oversight and undertaken in respect of fundamental rights, without undermining or circumventing the end-to-end encryption and confidentiality of the communications of all users by automatically and generally scanning encrypted communications; highlights that weakening encryption would undermine trust and security in digital services;

Added41. Highlights that the Member States’ legal systems do not keep up with the development and emergence of new types of CSAM, including synthetic, AI-generated CSAM as well as non-consensual intimate deepfakes, and calls for the criminalisation of the creation, possession and distribution of all types of CSAM regardless of the method of creation, and to require the swift takedown and cross-platform hashing of known abuse material, to stop it from reappearing;

Added42. Insists that tackling online sexual abuse and exploitation must also be done through prevention, reporting tools, education and appropriate law enforcement methods; urges the Commission and the Member States to ensure that victims have access to victim-centred and effective legal remedies; highlights the importance of accessible, confidential and child-friendly reporting channels that are based on safety and privacy by design, and that include specialised psychological and legal assistance for young victims and their parents; recalls that privacy is necessary for victims of abuse to protect their safety and increase their ability to seek help;

Added43. Highlights that all young people are affected by social media harms, including exposure to unrealistic body ideals, harassment, grooming and the non-consensual sharing of intimate material; underlines, however, that girls and young women are more exposed to social media harms, and are therefore disproportionately affected, particularly by body image pressure, harassment, misogynistic content and gender-based pressures, which can contribute to eating disorders, anorexia, self-harming behaviours and, in severe cases, extreme actions; stresses, therefore, the need for gender-sensitive digital safeguards to ensure safe, inclusive and empowering online environments for all young users; calls on online platforms to implement swift mechanisms for reporting and removing misogynistic content, to prevent algorithmic recommendations that amplify gender stereotypes, and to support positive campaigns on body image and gender equality;

Cyberbullying and online harassment

Change 8

Changed12.44. Stresses the importance of cyberbullying prevention and response tools, including awareness programmes, educational resources for teachers and families, and peer support structures, such as anti-bullying programmes; highlights the need to treat online and offline bullying holistically, recognising their interconnected nature;

Change 9

RemovedInfluencers and digital culture

Added45. Stresses the importance of understanding the phenomenon of the online disinhibition effect, which refers to the reduction of social restraints and empathy in online communication due to anonymity, physical distance and the absence of non-verbal cues; notes that this effect can lead to increased aggression, insults and cyberbullying; highlights the need to provide targeted and, where possible, individually tailored support for young people in school, at home, in institutional care and in other communal and social environments;

Removed13. Stresses the need for full transparency in influencer advertising that targets minors; underlines the importance of sponsorship disclosures and ethical standards for influencers targeting children, highlighting the risk of misleading marketing when advertisements are hidden as organic content;

Added46. Urges online platforms to ensure a swift and proportionate response to cyberbullying-related content and conduct, and to transparently report on its prevalence, and the response times and enforcement actions taken in countering it; stresses that content moderators play a crucial role in addressing online hate, and encourages online platforms to engage with independent civil society experts and researchers in developing and delivering their training, in this regard; underlines that transparency is essential to reveal the true extent of online hate, to identify potential algorithmic biases in detection systems and to ensure platforms’ accountability; calls on online platforms to cooperate with civil society organisations, educational institutions, child protection services, parents and caregivers in developing effective cyberbullying-prevention campaigns and in raising awareness of existing reporting mechanisms;

Added47. Calls on the Commission and the Member States to ensure long-term and sustainable funding, as well as systematic national visibility, for helplines and hotlines operating within the Safer Internet Centres network; stresses that these are key tools for providing immediate, professional and anonymous support to children and parents facing abuse or exposure to inappropriate online content;

AddedAction plan against cyberbullying

Added48. Welcomes the Commission’s action plan against cyberbullying and its common definition of cyberbullying at EU level; calls on the Commission and the Member States to ensure that online platforms are held responsible for preventing and countering cyberbullying, and to establish a European mechanism for the exchange of best practices among the Member States; underlines the importance of multidisciplinary cooperation between educational institutions, law enforcement authorities and relevant child protection services in the prevention and early detection of cases of cyberbullying; calls on the Commission to require online platforms to provide minors with an accessible reporting interface, and to ensure that platform design is safe by design and by default, including the possibility to block and report harmful contact and conduct;

AddedMental health

Added49. Stresses that social media can cause addiction and thus serious mental health harms, particularly for minors, with links to anxiety, depression, low self-esteem and body image issues;

Added50. Encourages awareness-raising campaigns targeting young people, families and schools, similar to initiatives such as Safer Internet Day, to prevent digital addiction, promote critical thinking, foster peer support skills and encourage healthier and more responsible social media habits; urges the Commission to develop comprehensive, evidence-based guidelines on healthy technology use and appropriate digital content, tailored by age group, focusing on usage patterns and screen time; stresses that such guidelines should serve as a reference for the design and safety features of digital services accessible to minors; calls on the Commission, in coordination with the Member States, to encourage online platforms to implement default user-friendly measures for minors’ accounts, including periodic prompts to take a break, configurable time-outs, and regular, unobtrusive notifications to users indicating the duration of their activity;

Added51. Highlights the positive role of sport and physical activity and community engagement in promoting the mental and physical well-being of young people; encourages the Commission and the Member States to promote regular offline activities, such as community-based arts, sports and civic engagement, as part of a broader strategy to reduce the excessive use of social media among young people and support their healthy development;

Added52. Calls on the Member States to ensure rapid access to mental-health support and psychological assistance for all young people, including by investing in grassroots, peer-led and community-driven psychosocial services; calls for psychological counselling in schools, dedicated support lines and awareness campaigns; calls for the implementation of suicide-prevention protocols on major online platforms, including the visible display of helpline numbers and support resources; warns that online tools, such as AI companions and chatbots, cannot replace mental health professionals and may be dangerous if used as a means of solving serious mental health issues;

Added53. Calls on online platforms to mitigate the impact of addictive and persuasive design features on the well-being of children and young people by implementing risk-based safeguards for recommender systems, and measures to effectively address, phase out and ban the most harmful addictive practices;

Added54. Calls for accessible and child-appropriate support services, including counselling, reporting mechanisms and legal assistance, for minors affected by AI-generated sexualised images and deepfakes;

AddedInfluencers and targeted advertising

Added55. Highlights the fact that social media influencers play an increasingly significant role in shaping the perceptions, aspirations and social attitudes of children and young people; welcomes the Commission’s efforts to ensure influencers’ compliance with existing legislation, including by setting up the Influencer Legal Hub; calls on the Commission and the Member States to develop a code of conduct for influencers, agencies and online platforms at EU level, including commitments to advertise responsibly towards minors, the clarification of responsibilities at different points of the value chain – such agencies representing influencers, and the brands whose products and services are being promoted – and the avoidance of harmful or misleading content;

Added56. Stresses the need for legislative action, including by establishing a harmonised definition of ‘influencer marketing’, and to strengthen, where appropriate, the prohibition of unfair commercial practices, in order to ensure full transparency in influencer advertising and influencer marketing that targets young people, as they have particular difficulty distinguishing between commercial and non-commercial content; underlines the importance of sponsorship disclosures and ethical standards, and highlights the risks presented by misleading marketing whereby advertisements are hidden as organic content, such as the consumption of unhealthy foods or engagement in gambling-like practices;

Added57. Calls on the Commission to protect minors from being commercially exploited in the context of ‘sharenting’ and ‘kidfluencing’ activities; calls on the Commission to take into account the progress made by some of the Member States in addressing sharenting, and to consider measures to limit this practice in order to protect children’s rights to their own images and to prevent their digital overexposure;

AddedHarmful content, age verification and the protection of minors

Added58. Regrets the fact that several mechanisms used by online platforms to control access to their services, such as self-declaration, are easily bypassed, and that minors still have widespread access to age-inappropriate content;

Added59. Stresses the need for effective, privacy-preserving age and child-rights compliant verification tools to verify a user’s age when required, without the collection or retention of personal data, in line with EU data protection standards, the AVMSD and the DSA;

Added60. Welcomes the development of the EU Digital Identity Wallet framework, which sets the standards for age-verification features; underlines that any age verification tool must strictly comply with the GDPR principles of data minimisation and purpose limitation, and should enable proof-of-age without revealing identity; supports models using trusted third parties with double anonymity mechanisms, in which the verifier cannot identify the service, and the service cannot identify the user, as provided for under the Regulation on electronic identification and trust services for electronic transactions in the internal market (which established the electronic Identification, Authentication and Trust Services (eIDAS) framework); calls on the Commission and the Council to promote coordinated investments and incentives for the deployment of the age-verification app across the Member States;

Added61. Calls, therefore, on the Commission to ensure that any age assurance mechanisms are supervised by public authorities, and comply with the privacy-preserving, non-intrusive and non-discriminatory-by-design architecture, interoperability and security requirements of the EU Digital Identity Wallet;

Added62. Calls on the Commission to establish, in coordination with the Member States, an EU mechanism for monitoring the effectiveness of child online protection policies, including the implementation of age restrictions, age-verification tools and safety settings on platforms, making use of existing capacities, in particular Eurostat data and methodologies; notes the current fragmentation across the Member States, and the fact that some of the Member States have implemented advanced measures to enhance the protection of minors online; encourages strong coordination at EU level, in this regard;

AddedMinimum digital age

Added63. Welcomes the favourable consideration expressed by Commission President Ursula von der Leyen regarding the introduction of a harmonised, EU-wide minimum age limit for access to social media (a minimum digital age); takes note of the initiatives already undertaken by certain Member States, in this regard; stresses the need for a harmonised EU approach to establishing a minimum digital age and, in this context, welcomes the conclusions of the Commission’s special panel of experts on child safety online, which provides guidance on this matter; welcomes, furthermore, the announcement made by President von der Leyen about a legislative proposal on the matter;

Added64. Recalls that Article 8 GDPR establishes age thresholds for valid consent between the ages of 13 and 16, which can be determined nationally; underlines the lack of implementation of Article 8(2) GDPR, which stipulates that the controller should make reasonable efforts to verify that consent is given or authorised by minors’ guardians; stresses that existing instruments should be duly implemented by service providers, including privacy-by-design and privacy-by-default, data minimisation and purpose-limitation principles; recalls that the Commission guidelines under Article 28 of the DSA emphasise that access restrictions and age-assurance mechanisms alone cannot be substitutes for other, more structural measures recommended in the guidelines; stresses that Article 28b(3), point (f) AVMSD provides for the establishment of age-verification systems on VSPs to protect minors from content that may impair their physical, mental or moral development; underlines that the effective implementation of such systems should lead to a safer online environment for children and young people, while fully respecting their privacy;

AddedParental control

Added65. Commends existing parental control tools and calls on the Commission to develop a non-prescriptive, technology-neutral EU framework of minimum functional standards for such tools, accompanied by an EU certification scheme to increase uptake and cross-device effectiveness; stresses that these tools must be user-friendly, privacy-respecting by default and accessible to all parents and guardians;

Added66. Expresses concern that parents and guardians often lack the knowledge or tools to effectively monitor their children’s online activities, and that even when parental controls exist, they are frequently difficult to find or easily circumvented; urges platforms to take greater responsibility for the effectiveness and visibility of these systems, and calls on them to publish data on their uptake and performance;

Added67. Urges online platforms to take greater responsibility for protecting children and young people from harmful or manipulative content; encourages, furthermore, the Commission to consider introducing personal liability in cases of serious and persistent non-compliance with the protection provisions concerning minors set out in Articles 34 and 35 DSA, in order to incentivise better compliance;

Added68. Underlines that regulated media service providers, in particular independent public service media providers regulated in line with the principles of the European Media Freedom Act, offer a broad range of content and services for children and young people that are adapted to their development needs; calls on online platforms to ensure their recommender systems and content moderation mechanisms give more visibility to reliable and age-appropriate content and services provided by such professional media organisations;

Added69. Calls on the Commission, in cooperation with the Member States, relevant bodies, platforms and experts, to develop specific guidelines and tools to combat the spread of disinformation targeting young people online; stresses the need to integrate the labelling of suspicious content, the removal of false information that may endanger children’s health or safety, and the promotion of reliable sources through educational tools and algorithmic recommendations;

AddedSafeguarding tools for young people in online platform designs