Drawing the Line: protecting people without silencing expression

When online safety rules treat fiction, queer expression and survivors’ testimony as forms of abuse, who benefits, and who loses protection? This question brought together participants in Kathmandu and online for the Center for Online Safety and Liberty’s workshop at the Asia Pacific Regional Internet Governance Forum (APrIGF) on 6 October 2026.

“Drawing the Line: Online Harm, Queer Expression, and Platform Governance in Asia-Pacific” examined how laws and platform policies can distinguish personal expression from abuse involving real people. Jeremy Malcolm, Shambhawi Paudel and KS Park approached that distinction through research, regional community experiences and South Korean law. Paudel also shared observations supplied by Mish Pony, who was unable to attend.

Jeremy Malcolm: why the distinction matters

COSL Chair Jeremy Malcolm opened by introducing the Drawing the Line Principles and Watchlist. The Principles provide a framework for protecting expression while addressing abuse, including respect for survivor autonomy, evidence-based prevention and meaningful digital due process. The Watchlist examines how national laws observe or erase these distinctions, concentrating on the treatment of fictional material under child sexual abuse laws.

Malcolm distinguished images recording the abuse of a real child, manipulated images targeting an identifiable child, and wholly fictional material involving no real child. The first two can directly violate a person’s rights. The third requires a separate account of the harm alleged and the justification for restricting it. Applying the same terminology and criminal framework to all three obscures that difference.

These questions reach beyond pornography. Graphic memoirs and literature can depict sexuality or abuse to explore identity, recount experience or give survivors a voice. Malcolm discussed works including Gender Queer and A Girl on the Shore, warning that the presence of a sexual depiction alone tells us little about a work’s purpose or its effects. Rules ostensibly protecting survivors can end up suppressing their accounts.

His Australian example was Lauren Mastrosa’s prosecution over the novel Daddy’s Little Toy. As Malcolm explained, its central relationship involves adults, with one character role-playing as a child. Nevertheless, the author was convicted of child abuse material offences. For Malcolm, the case illustrated the consequences of a legal category that encompasses both fictional writing and recordings of actual abuse.

The distinction also matters for enforcement. Drawing on the Watchlist’s UK research, Malcolm described rising prosecutions involving fictional material alongside falling prosecutions involving conventional CSAM. He acknowledged that these trends do not establish a one-for-one displacement of cases, while questioning how enforcement priorities serve children when increasingly directed towards fiction.

Malcolm also challenged assumptions that consuming fictional depictions necessarily causes offending. He described research identifying possible risks, possible protective uses and substantial gaps in knowledge. Those uncertainties support further research and access to clinical assistance; they do not establish a simple causal justification for blanket criminalisation.

He proposed proportionate responses suited to particular harms: user controls and content labels, limits on recommendations, confidential support, and targeted action against grooming or harassment. Clear moderation reasons and meaningful appeals would help platforms address harmful conduct while protecting legitimate expression.

Shambhawi Paudel, with contributions from Mish Pony: whose expression disappears?

Shambhawi Paudel of ILGA Asia examined what happens when governments and platforms draw these boundaries through vague ideas of morality and obscenity. Queer expression, she explained, can be suppressed through broadcasting rules, policing and platform governance even without an explicit prohibition on LGBTIQ content.

Her regional examples showed how fiction and everyday visibility become targets. In China, she discussed arrests of writers of danmei, or boys’ love fiction, including young women publishing online. She also highlighted the reported detention of two men over an AI-generated image of pandas engaged in same-sex behaviour. Such cases exposed the breadth of censorship justified through sexual morality.

In Malaysia, Paudel described restrictions on queer publications and dating services alongside official efforts to replace the term “LGBT” with “deviant culture”. She drew on research by Justice for Sisters and Amnesty International Malaysiadocumenting censorship of LGBTIQ publications. Her examples extended to the seizure of rainbow-themed Swatch watches, illustrating how even symbols of visibility can become contested objects of regulation.

Paudel connected these restrictions to threats outside formal legal processes. She described online vigilante campaigns targeting queer people and the use of dating platforms for entrapment, including examples documented in the Middle East and North Africa. For people facing detention or violence, anonymity and careful control over disclosure can be essential safeguards. Measures that remove those safeguards may increase their exposure to harm.

Turning to Indonesia, she discussed the classification of LGBTIQ issues as a non-military threat and proposed broadcasting restrictions targeting LGBT content. Across these examples, her concern was a pattern of shrinking space for expression and association, reinforced by discriminatory laws and hostility online.

Platforms can reproduce the same exclusions. Paudel pointed to the effects of Tumblr’s adult-content ban and automated moderation that treats queer material as inherently sexual. Health information, discussions of identity and survivors’ accounts can disappear alongside the content a platform intended to restrict. This makes it harder for communities both to express themselves and to document abuse.

Paudel then relayed observations supplied by Mish Pony, former CEO of Scarlet Alliance and CEO of Flat Out. Pony had highlighted Australian consideration of a ban on depictions of choking in pornography, following the UK, and a proposal to allow the Minister for Communications to designate and regulate new “online harms”. Their notes questioned the recurring assumption that representation and fictional depictions necessarily lead to enactment in real life.

Pony also drew attention to overlapping anti-abortion, anti-LGBTI and conservative Christian advocacy networks. They warned that these actors could use genuine concerns about violence against women and online harassment to advance agendas restricting rights. In discussing the Australian context, Paudel additionally raised eSafety’s attempted global takedown of footage of a church stabbing and the limits of regulatory jurisdiction.

Together, these contributions asked whose experiences shape definitions of harm, and whose voices those definitions exclude. Safety policies need to account for the risks created by censorship itself, especially for communities already exposed to discrimination.

KS Park: when child protection law loses sight of children

Joining remotely, KS Park of Korea University and Open Net Korea described a “misalignment” between the purpose of South Korean child protection law and its application to fictional depictions.

Park traced the expansion of the law to cover material featuring characters perceived as children. His central concern was that severe penalties developed to address exploitation of real children were being applied to cartoons and animation. He contrasted the treatment of these depictions with ordinary obscenity offences, questioning the proportionality of the resulting punishments and associated restrictions.

The rationale for treating CSAM as an exceptional category, Park explained, rests on the abuse of a child and the continuing harms associated with circulation. Images manipulated to target an identifiable child can also cause direct harm. Wholly imaginary characters raise a different question: which person is the law protecting, and through what mechanism?

Park challenged both the claim that fictional depictions stimulate abuse and the appeal to sexual morality as a sufficient justification. He used American Beauty and Traffic to illustrate the breadth of sexual themes involving fictional young people in acclaimed cinema. Depiction alone cannot resolve whether a work should attract criminal punishment.

He also argued that conflating categories could undermine child protection in practice. Police incentives might encourage pursuit of easily located fictional material, while young consumers of cartoons could themselves become targets of investigation. He further suggested that combining cases involving real and fictional depictions could distort sentencing data, potentially affecting judicial expectations in cases involving actual victims. These were reasons, in his analysis, to examine the institutional effects of the law alongside its stated intentions.

Park discussed separate legal treatment of fictional material as one possible reform, while making clear that he was not endorsing its criminalisation. His priority was to restore attention and judicial resources to protecting real children.

Q&A: context, accountability and participation

A participant from Taiwan described controversy over the treatment of fictional characters under child protection rules, asking whether criminalisation required a real victim and whether age classification could offer a better response. Park reiterated his concern about expanding child exploitation laws to cover images without a real child and offered to engage further with Taiwanese advocates.

Paudel stressed the importance of context: what an image communicates, whether it targets a real person or solicits harmful action, and who uses the platform where it appears. A policy response should also consider whether it creates a precedent that extends censorship without improving protection.

A participant from South Korea supported keeping enforcement focused on harm to living children while protecting queer expression and survivors’ stories. The Taipei local hub described concerns about platform obligations and blocking powers, calling for meaningful participation by artists, writers, parents and platforms in policymaking.

Time prevented a full response to a question about enforcement against online platforms in Korea and further contributions from remote participants. Even within that constraint, the discussion reinforced a shared demand: online safety governance should identify the harm it seeks to prevent, explain how its interventions address that harm, and remain accountable to the people whose rights it affects.

Watch the workshop recording.