About this article
The title uses the word “experiment”, but no experiment was run and no data of our own is reported. This is a general explainer on online harassment: what it looks like, why it is hard to stop, and general steps people can take. It is not legal, medical or psychological advice. Anyone in danger should contact local emergency services, and anyone who is distressed should consider speaking with a qualified professional or a trusted adult.
What counts as cyberbullying
Cyberbullying is generally understood as repeated hostile behavior carried out through digital channels, intended to hurt, embarrass or frighten someone. The term is often used for young people, while “online harassment” is the broader label that also covers adults. Definitions vary between countries, schools and platforms, and the legal treatment differs from place to place.
Common forms include:
- Abusive messages and comments, public or private, including insults and threats.
- Exclusion, such as removing someone from group chats or deliberately leaving them out of shared online spaces.
- Spreading rumors or private information, sometimes called doxxing when it involves home addresses or other identifying details.
- Impersonation, where someone creates a fake account or hijacks a real one to damage a person’s reputation.
- Sharing embarrassing or intimate images without consent.
- Pile-ons, in which many people direct hostile messages at one target within a short time.
- Persistent unwanted contact, often called online stalking.
- Manipulated media, such as edited images or fabricated clips. Advances in deepfake technology mean that fake material can be harder to recognize, which adds to the harm for those targeted.
Not every online conflict is bullying. A single rude comment or a disagreement between people of equal standing is different from a pattern of targeting. That said, a one-off incident can still be harmful and worth reporting, especially if it involves threats or private images.
Why it is hard to stop
Constant access. In earlier eras, a bullied child could often get relief at home. Phones and messaging mean that unwanted contact can follow someone anywhere, at any hour.
Permanence and reach. Posts can be copied, screenshotted and re-shared. Even after something is deleted from one place, it may persist elsewhere, and an audience can grow far beyond the original group.
Anonymity and distance. Fake or throwaway accounts can make it easier for people to act in ways they would not in person. Not seeing the other person’s reaction may also reduce empathy. Researchers have discussed this “online disinhibition” idea, though it does not explain all harassment, and many people are harassed by people they know.
Scale and moderation limits. Platforms receive enormous volumes of content. Automated systems can catch some abuse, but they make mistakes in both directions, and bias in algorithms can affect what gets flagged. Context, slang and sarcasm are difficult for software, and human review is limited by cost and speed.
Differing rules. Each platform defines harassment differently, and an offender who is blocked in one place may move to another. Laws also vary and may not clearly cover every behavior.
Bystanders. Onlookers may not know what to do, may fear becoming the next target or may assume someone else has already reported it. Group dynamics can also reward cruelty with attention.
Effects on those targeted
Experiences vary. Some people recover quickly, while others report anxiety, withdrawal, trouble sleeping, difficulty concentrating or low mood. Research on how serious and lasting the effects are is mixed, partly because studies use different definitions, but there is broad agreement that harassment can be distressing and that serious cases deserve prompt support.
General steps if it happens
These are common suggestions, not guarantees, and the right response depends on the situation.
- Do not feel it is your fault. Responsibility lies with the person doing the harassing.
- Save evidence. Take screenshots with dates, usernames and links before anything is deleted. Keep them in a safe place.
- Avoid retaliating. Replying angrily can escalate things, though staying silent is not a requirement and the choice is yours.
- Use block and mute tools. Most platforms let you block accounts, limit who can message or comment, and filter words.
- Report the content. Use the platform’s reporting feature and select the closest category. Reports are often reviewed more easily when evidence is clear.
- Review your privacy settings. Limiting who can see or contact you may reduce further exposure, although it cannot undo what has already been shared. Basic cyber safety habits, such as strong passwords and two-factor authentication, also help protect accounts from being hijacked.
- Tell someone you trust. This might be a parent, friend, teacher, workplace manager or counselor.
- Escalate when needed. Threats of violence, sharing of intimate images, or stalking may be matters for the police or other authorities. Schools and employers often have their own procedures.
- Look for support. Helplines and counseling services exist in many places, and a doctor or therapist can help if the experience is affecting your wellbeing.
For parents and carers
Young people sometimes hide online problems because they fear losing access to their devices. Calm, open conversation tends to work better than punishment. It helps to explain that they will not be blamed, to learn how the reporting tools work, and to agree on what to do if something goes wrong. Families may also want to read about children online privacy.
For bystanders
Simple actions can matter: do not share or reward hostile posts, report them, and check in privately with the person targeted. Public confrontation is not always safe or helpful, so choose carefully.
Open questions
How much responsibility should platforms carry, and how should laws balance protection from abuse with freedom of expression? These are debated, and approaches to AI regulation and platform rules continue to change. What does seem clear is that no single tool solves the problem, and that a mix of better design, clear rules, education and human support is usually discussed as the way forward.