Areas of Harmful Online Content

The first area pertains to the threats that challenge information authenticity and integrity, which arise from misinformation and covers harmful online content that is not based on facts, including deepfakes. A1. Misinformation Misinformation refers to a broad coverage of harmful online content not based on fact. Misinformation involves alteration, missing context, and misrepresentation. Based on the World Economic Forum (WEF) typology [1], misinformation involves “the dissemination of incorrect facts, where individuals may unknowingly share or believe false information without the intent to mislead.” A2. Disinformation Disinformation involves the “deliberate and intentional spread of false information with the aim of misleading others” [1]. Disinformation can be understood as a specific type of misinformation with an intentional element. A3. Deceptive synthetic media Deceptive synthetic media refers to “content that has been generated or manipulated via algorithmic processes (such as artificial intelligence or machine learning) to appear as though based on reality, when it is, in fact, artificial and seeks to harm a particular person or group of people.” [1] Deepfakes (highly realistic, though fake, image, audio or video content) are one type of such deceptive synthetic media.

The vast majority of social media content contains opinions, feelings, and comments that do not pertain to facts. While some of such content fosters connection, some of these subjective contents contain hateful content, disrespectful language or harassment, which cause lasting negative emotional impacts on our population. This area pertains to threats that challenge dignity and social well-being of individuals and communities. B1. Toxicity Toxic content broadly refers to disrespectful content. It can range from a slightly dismissive comment to the use of very offensive or abusive language. Hate speech is a specific type of online toxic content. B2. Online harassment and bullying Online bullying and harassment content can be hostile language/image/video, hurtful content, or content that socially isolates or excludes a person. B3. Hate speech Hate speech refers to “Any kind of communication in speech, writing or behaviour that attacks or uses pejorative or discriminatory language with reference to a person or a group on the basis of their inherent/protected characteristics – in other words, based on their religion, ethnicity, nationality, race, colour, ancestry, gender or other identity factor.” [1]. Specific hate speech can take on the forms of racism, misogyny, sexism, xenophobia, homophobia, etc. Contexts, such as cultural, linguistic, and content modality, often make determining or detecting hate speech a highly complex task. Related to this broad area, it is useful to note that regulations may use more outcome-based definitions. In the Broadcasting Act Part 10A, one type of egregious content is content likely to cause racial and religious disharmony in Singapore, which refers to “content dealing with matters of race or religion in a way that is likely to cause feelings of enmity, hatred, ill will or hostility against, or contempt for or ridicule of, different racial or religious groups in Singapore.” [2]

This area pertains to threats to health and well-being, which often involves content that is dangerous or inappropriate, especially for vulnerable groups (e.g., children). C1. Content that promotes self-harm This refers to content that “promotes, encourages, coordinates or provides instructions on: suicide, Self-injury, or Eating disorders” [1]. C2. Content that is developmentally inappropriate This refers to content that includes “children’s access to pornography, particularly of a violent or extreme nature, and graphic, violent material.” [1]

This area includes seven content risks, including child sexual abuse material (CSAM), child sexual exploitation material (CSEM), pro-terror material, content that praises, promotes, glorifies or supports extremist organisations or individuals, violent graphic content, content that incites, promotes or facilitates violence, and content that promotes, incites or instructs in dangerous physical behaviour [1].

1. World Economic Forum – Typology of Online Harms. https://www3.weforum.org/docs/WEF_Typology_of_Online_Harms_2023.pdf 2. Singapore Broadcasting Act 1994 – Part 10A Online Communication Service Regulation - Section 45D Meaning of “egregious content”. Singapore Government Website, https://sso.agc.gov.sg/Act/BA1994?ProvIds=P110A-#pr45D-s