Stalking and harassment (overview)

AI tools can make harassment worse—cheaper messages at scale, fake profiles, or nonconsensual imagery (see that chapter). This page explains risk patterns victims and platforms discuss publicly, not how to harass anyone.

Safety literacy helps people recognize abuse, support targets, and design products with fewer sharp edges.

• Automated messaging and deepfake-enabled abuse reported in news and advocacy work.
• Platforms publish policies against stalking/harassment; enforcement is imperfect.
• Workplace and school policies increasingly cover AI-assisted abuse.

• Do not use AI to contact, track, intimidate, or impersonate people without consent.
• If you are targeted: document, block, use platform reporting, and contact local authorities when threatened.
• Builders: rate-limits, abuse detection, and human review matter.

• No tracking methods, no evasion of blocks, no social-engineering scripts, no code.

Local emergency services for immediate danger. National domestic-violence / cyberstalking hotlines in your country (look up current official numbers—do not rely on a chatbot alone).

⚠️ Myth: “It’s just a joke prompt.”

✓ Reality: Impact on the target defines harm; intent is not a free pass.

⚠️ Myth: Public profiles mean anything goes.

✓ Reality: Consent and law still apply.

• Platform safety policy hubs (cite the ones you verify)
• Course 29: nonconsensual-imagery; scams-and-fraud
• NIST AI RMF (abuse/misuse framing): https://www.nist.gov/itl/ai-risk-management-framework ↗