Trust & Safety: The Stakeholders of Digital Trust Under Pressure

Difficult adoption of the DSA, challenges to moderation policies, increase in toxic online content... The world of digital security and trust is going through a turbulent period. The latest edition of the Trust & Safety Forum provided an opportunity to take stock.

On April 1st and 2nd, the Trust & Safety Forum was held in Lille, alongside the InCyber Forum (FIC). For its fourth edition, this international event dedicated to promoting a safer and more trustworthy digital environment brought together all Trust & Safety (T&S) stakeholders: digital platforms, regulators, NGOs, and trusted flaggers.

The Forum first provided an initial assessment of the implementation of the Digital Services Act (DSA). This European Union regulation on digital services aims to combat the spread of illegal content online. Initially targeting very large online platforms (VLOPs), since February 17, 2024, it applies to all actors operating in the European market.

Among other obligations, platforms must regularly publish transparency reports detailing their moderation policies with supporting data. To harmonize these reports and facilitate their comparison, the European Commission defined a standardized data model last November. From July 2025, platforms will be required to collect their data according to these “templates.” The use of a unique “taxonomy” will ensure consistency of the reported data. The first standardized transparency reports are expected to be published by February 2026.

This change significantly impacts platform workloads. Louis-Victor de Franssu, co-founder of Tremau, a company specializing in platform compliance with the DSA, notes that extracting certain data from existing systems can take three weeks to a month. “Furthermore, the proliferation of tools — including those not dedicated to trust and safety, like Salesforce or Zendesk — complicates data collection even more.” Standardization is all the more critical as large platforms must also produce reports for other regulators, such as the UK’s Online Safety Bill or Australia’s Children’s Online Privacy Code.

Samuel Comblez, Deputy General Director of the association e-Enfance, agrees. “The greatest challenge is accessing reliable information. For example, our current software cannot precisely track the time it takes a platform to respond. This is crucial since the association received 160,000 requests in 2024.”

Margaux Liquard, Trust and Safety Lead at Yubo, a social app for teenagers, says she strives to meet trusted flaggers’ expectations by consolidating requested information and offering a dedicated email address. She views this commitment to trusted actors as a competitive advantage, making the platform safer for its users.

Moderation Policies Under Strain

While the DSA serves as a safeguard in Europe, the situation across the Atlantic leans towards deregulation and challenges to traditional moderation policies since Donald Trump’s return to power. In January, Meta decided, following X’s example, to stop using human fact-checkers and rely on AI and user self-moderation for content on Facebook and Instagram.

The end of human fact-checking raises fears of an increase in violent and toxic discourse online. Though initially not affecting Europe, this decision appears to have extended to the continent. In Spain, over 2,000 moderators working for a Meta subcontractor were recently laid off.

Henry Adams, Trust & Safety Director at Resolver, notes this trend predates Meta’s policy shift. He points out that the “Overton window” — the range of ideas and practices considered acceptable by society — has widened over the years. “Content once deemed extreme is now commonplace.”

Dr. Beatriz Lopes Buarque from the London School of Economics and Political Science studies conspiracy theories linked to the far right. She observes that such theories circulate widely on digital platforms and are legitimized by conferences, documentaries featuring intellectuals, and academic works indexed by Google Scholar.

She advocates for an additional button on social media to report problematic content alongside “like,” share, or comment options.

She also cited Netflix’s series Adolescence, which depicts a teenager gradually influenced by British influencer Andrew Tate’s masculinist ideology. According to the “red pill” concept — a Matrix-inspired metaphor for awakening — teens access legally permissible but harmful content that serves as a gateway to extreme and hateful theories.

For Lopes Buarque, the problem cannot be solved by regulation and moderation alone. “It’s much more serious.” She calls for a civil society awakening to protect children and praises alternative platforms like BlueSky — a social network founded in 2019 by Jack Dorsey — which offer relatively safer spaces for discussion.

Adolescents Highly Exposed to Toxic Content

Vaishnavi J., founder of Vyanams Strategies (VYS), helps platforms, civil society, and governments implement strategies to safeguard children online.

She studied gender-based behaviors in video game players. While young men are the majority, moderation tools target all players, including young women and non-binary teens. Notably, the latter groups are more likely to report inappropriate behavior.

“At first glance, it may seem young women have a rough time gaming, but boys can also be victims of abuse — they just don’t speak up,” she explains. She refers to the Digital Thriving Playbook, a resource helping developers create safer, more inclusive games.

Instead of viewing T&S departments as cost centers, platforms should promote their moderation policies. Offering “safe” spaces, social networks would gain or at least retain audiences, she argues.

Vaishnavi J. notes many youths who experienced online harassment have turned away from traditional social media, preferring video games or private messaging to interact with friends.

In this context, regulators and states can pressure platforms. The French government is considering banning social media access for minors under 15, requiring platforms to verify age.

Iain Corby, director of the Age Verification Providers Association (AVPA) in London, says solutions exist on Google Wallet and Apple’s App Store but remain imperfect. Children can circumvent controls by using parents’ IDs.

Simon Newman, CEO of the Online Dating and Discovery Association (ODDA), highlights the costs of age verification — a few cents per operation — which may deter small free platforms like dating apps.

He is skeptical about facial recognition for age estimation, as planned in Australia, where a law banning under-16s from social media takes effect in November 2025. “A 15-year-old could bypass facial recognition, while a 17-year-old might be blocked on TikTok. There will be riots in schoolyards!”

Trusted Flaggers: An Exposed Status

A roundtable focused on the delicate position of trusted flaggers. Facing increasing cyberattacks and disinformation campaigns, they also suffer financial pressures limiting their capacity.

While trusted flaggers and whistleblowers are prime cybercrime targets, Kayle Giroud, director of commons initiatives at Global Cyber Alliance, highlighted the organization’s toolkits and resources to protect associations and NGOs.

In France, the Audiovisual and Digital Communication Regulatory Authority (Arcom) grants legitimacy to trusted flaggers. Recently, it designated seven associations, including e-Enfance and Point de Contact, as trusted flaggers responsible for reporting manifestly illegal content to platforms.

“The trusted flagger’s role bridges citizens and platforms,” explains Alejandra Mariscal, director of Point de Contact. “We cannot remove content ourselves — we notify platforms, who decide whether to remove it.”

She warns against narratives equating regulation with censorship, which fuel hostility against trusted flaggers. “This can lead to violence or cyberattacks if people see us as enemies of free speech,” the exact opposite of their mission.