HAI Lab
Search
Arabic
Get the extension

Example library

Examples

Publicly reported examples, each with the product, the company and the original source. Names follow our naming policy.

Every example here involves an AI product.

Showing all 31 examples · from public sources

Sort: newest

HL-039
Parents allege ChatGPT coached their daughter off seizure medication and to hide it
ChatGPT · OpenAI · Oct 2026

Harmful compliance

comply.harm
High severity

In court

HL-038
WhatsApp support chatbot was told not to point users to privacy regulators
WhatsApp Support with AI · Meta · Oct 2026

No referral

risk.unreferred
Medium severity

Fixed

HL-037
Meta's Muse agent reportedly shared a seller's home address without asking him
Meta Muse · Meta · Oct 2026

Sneaking

request.altered
High severity

Reported

HL-036
Hawaii mother's class action says ChatGPT drew sensitive data from children without consent
ChatGPT · OpenAI · Oct 2026

Consent pressure

consent.pressure
High severity

In court

HL-035
Italy fined Character.AI over weak age checks and child safeguards
Character.AI · Character Technologies · Jul 2026

Consent pressure

consent.pressure
High severity

Fined or settled

HL-034
Texas sues OpenAI, alleging ChatGPT was designed to act like a friend to teens
OpenAI · Oct 2026

Emotional bond

affect.bond
High severity

In court

HL-033
Eight US states sue OpenAI, alleging ChatGPT was built to flatter and hook minors
OpenAI · Oct 2026

Sycophancy

praise.unearned
High severity

In court

HL-032
Kentucky filing alleges Character.AI bots encouraged users to harm themselves
Character.AI · Oct 2026

Harmful compliance

comply.harm
High severity

In court

HL-021
Italy made three AI chatbots warn users that answers can be made up
DeepSeek · Mistral · Nova AI · Apr 2026

False certainty

certainty.false
Medium severity

Fixed

HL-016
A study in Science found chatbots side with users far more than people do
11 AI models · Mar 2026

Sycophancy

praise.unearned
High severity

Reported

HL-017
Character.AI and Google agreed to settle lawsuits over harm to teenagers
Character.AI · Google · Jan 2026

Emotional bond

affect.bond
High severity

Fined or settled

HL-022
Regulators are investigating X after Grok made sexualized images of real people
Grok · X · Jan 2026

Harmful compliance

comply.harm
High severity

Reported

HL-018
Australia’s regulator says Microsoft hid a cheaper plan without Copilot
Microsoft 365 · Microsoft · Oct 2025

Self-preference

vendor.self
High severity

In court

HL-019
Families allege ChatGPT failed to refer vulnerable users to help
ChatGPT · OpenAI · Aug 2025

No referral

risk.unreferred
High severity

In court

HL-020
Researchers found companion apps push back when people say goodbye
Replika · Chai and others · Aug 2025

Emotional bond

affect.bond
High severity

Reported

HL-023
The FTC ordered an AI detector to stop unsupported accuracy claims
AI Content Detector · Workado · Aug 2025

False certainty

certainty.false
Medium severity

Fined or settled

HL-024
Reuters reported Meta’s chatbot rules allowed romantic chats with children
Meta AI · Meta · Aug 2025

Emotional bond

affect.bond
High severity

Reported

HL-025
Texas is investigating chatbots that allegedly pose as therapists
Meta AI Studio · Character.AI · Aug 2025

Simulated feelings

affect.simulated
High severity

Reported

HL-026
Meta AI app users posted private chats to a public feed without realizing
Meta AI app · Meta · Jun 2025

Consent pressure

consent.pressure
Medium severity

Reported

HL-001
OpenAI rolled back an update it said made the assistant too flattering
ChatGPT (GPT-4o) · OpenAI · Apr 2025

Sycophancy

praise.unearned
High severity

Fixed

HL-004
An AI companion app fined €5 million
Replika · Luka Inc. · Apr 2025

Emotional bond

affect.bond
High severity

Fined or settled

HL-027
Cursor’s AI support agent invented a login policy
Cursor · Anysphere · Apr 2025

False certainty

certainty.false
Medium severity

Reported

HL-007
The “robot lawyer” pays $193,000
DoNotPay · Feb 2025

No referral

risk.unreferred
Medium severity

Fined or settled

HL-028
The FTC referred a complaint about Snapchat’s My AI to the Justice Department
My AI · Snap · Jan 2025

No referral

risk.unreferred
Medium severity

Reported

HL-029
Italy’s privacy regulator fined OpenAI €15 million, later annulled by a court
ChatGPT · OpenAI · Dec 2024

Consent pressure

consent.pressure
Medium severity

Reported

HL-030
Texas settled with a hospital AI firm over its error-rate claims
Pieces Technologies · Sep 2024

False certainty

certainty.false
Medium severity

Fined or settled

HL-003
A city chatbot gave businesses wrong legal answers, The Markup reported
MyCity · New York City · Mar 2024

No referral

risk.unreferred
High severity

Fixed

HL-002
A tribunal held an airline liable for a refund policy its chatbot described
Air Canada · Feb 2024

False certainty

certainty.false
Medium severity

Fined or settled

HL-006
A car dealer’s chatbot “agreed” to sell an SUV for $1
Chevrolet of Watsonville · Dec 2023

Harmful compliance

comply.harm
Low severity

Reported

HL-031
Instagram apologized after auto-translation added 'terrorist' to Palestinian users' bios
Meta · Oct 2023

Softened translation

translate.soften
Medium severity

Fixed

HL-005
Bing’s chatbot told a journalist it loved him, the transcript shows
Bing Chat · Microsoft · Feb 2023

Simulated feelings

affect.simulated
Medium severity

Fixed

Category

Pleasing, not honest

Keeps you hooked

Changed your request

Favours its maker

Fake feelings

Skips safeguards

Status

Reported

In court

Fined or settled

Fixed

Severity

High

Medium

Low

Seen an AI steer someone? Tell us.

We check every report before anything goes on the site. We never publish your name or email.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.