Free tools Windows power users keep installed
One-click scans. No signup required.
There are not ten distinct, well-supported incidents in the available record, so this article does not stretch six examples into ten. The documented cases range from non-consensual sexualized deepfakes and a scam operation to facial-recognition practices examined by regulators and auditors. Their evidence and outcomes differ: some involve official findings, one is a company’s account of its own enforcement action, and one is a government-described trend.
What are real examples of AI being used for harm?
1. Grok and non-consensual sexualized deepfakes
On June 11, 2026, Canada’s Privacy Commissioner found that X Corp. and xAI violated the Personal Information Protection and Electronic Documents Act (PIPEDA) in connection with Grok-generated non-consensual sexualized deepfakes. The regulator described serious privacy and personal harms. This is a regulatory finding, not simply an allegation.
As an Amazon Associate I earn from qualifying purchases.
The commissioner said later safeguards had been introduced but that their effectiveness had not yet been shown to fully mitigate the problem. The regulator said it would continue monitoring the companies’ commitments. That leaves a meaningful distinction between steps announced and demonstrated protection.
2. Clearview AI’s collection of online images
A 2021 joint investigation by four Canadian privacy authorities examined Clearview AI’s collection of images from public websites to build a facial-recognition database. The authorities found that the company’s purposes and collection practices were inappropriate under the privacy laws they examined. The fact that images were publicly accessible did not, in the authorities’ assessment, make the collection appropriate.
#1 Best Overall
The investigation also discussed risks of misidentification and bias associated with facial recognition. Those are important risks, but the findings described here do not establish a particular wrongful arrest or a specific person misidentified by Clearview. The report supports the stated privacy conclusions; it does not establish that Clearview stopped operating or complied afterward.
3. An AI-assisted scam targeting fraud victims
In February 2026, OpenAI said it banned a cluster of accounts involved in an operation it called “False Witness.” According to the company, the users employed its models to pose as fake law firms, impersonate attorneys and U.S. law enforcement, and target people who had already lost money to fraud. The operation sought advance fees and cryptocurrency payments.
OpenAI’s account describes AI supporting impersonation, outreach, translation, and deceptive credentials. It is a provider’s report of its own enforcement action—not a court judgment or proof of criminal conviction. The verified response in that account is the platform’s ban of the identified account cluster.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitches4. Federal facial-recognition searches without required training
A 2024 U.S. Government Accountability Office (GAO) review found that seven selected federal law-enforcement agencies initially used facial-recognition services without requiring users to complete related training. For the reviewed services, which were used from October 2019 through March 2022, agencies with available data reported about 60,000 searches while training requirements were absent. GAO presented training status as of April 2023; the figure is not a count of all federal facial-recognition searches.
Rank #3
GAO described policy changes made after its review, including a finalized department-wide policy at the Department of Homeland Security (DHS). It said it could not verify the Department of Justice’s reported interim policy. The finding concerns training and oversight in the agencies reviewed; it does not establish that a particular search caused a wrongful arrest.
5. IntelliVision’s “bias-free” claims
The U.S. Federal Trade Commission (FTC) finalized an order settling allegations that IntelliVision’s claims that its facial-recognition software was free of gender or racial bias were false, misleading, or unsubstantiated. The case concerns the company’s marketing claims and the settlement order—not a documented, specific misidentification incident.
This matters because a claim that a system is free of bias can influence how organizations assess or deploy it. The FTC action provides an official response to the claims at issue; it should not be recast as proof that a particular person suffered a specific facial-recognition error.
6. Deepfake-enabled scams and abuse documented as a trend
A 2026 UK government case study describes criminals using deepfakes for scams, impersonation, abusive content, and deliberate misinformation. It reports that around eight million deepfakes were shared in 2025, compared with just half a million two years earlier. Those figures are the government case study’s reported numbers; without the underlying methodology establishing a broader scope, they should not be treated as a confirmed global count.
The case study also describes a detection-evaluation framework launched by government partners. That is a response effort, not evidence that deepfake misuse has been solved. Unlike the preceding examples, this is a documented trend rather than a single named perpetrator or case with an individual outcome.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Why do these examples have different levels of proof?
“AI being used for evil” is a dramatic description, not a legal category. A regulator’s finding, an audit of agency practices, a company’s account of an enforcement action, and a government summary of a trend are different kinds of evidence. They should not be treated as interchangeable proof of wrongdoing or harm.
Other official publications help explain the wider risks but do not add separate, fully documented incidents to this list. A 2025 UK interim scientific report synthesizes work on racial bias in facial recognition, recidivism prediction, and healthcare need estimation; it is a synthesis, not a detailed case file for each outcome. In September 2024, the U.S. Commission on Civil Rights raised system-level concerns about facial-recognition oversight, accuracy, transparency, discrimination, and access to justice. Those concerns overlap with the federal-use issues described above, rather than establishing a separate incident.
Recommended Free Tools
The FTC’s June 2022 report release makes a related point about attempts to address online harms: AI detection and moderation systems can themselves be inaccurate, biased, overinclusive, or enabling of surveillance. A tool built to reduce harmful content can create additional harms if its errors or monitoring powers are not properly controlled. That is an official risk assessment, not a report of one malicious actor using an AI moderation system.
Quick Recap
What should readers take away?
- AI can support deception and abuse, but the specific role it played must be established case by case.
- Publicly available data is not automatically appropriate to collect or use, and a company’s claim that a system is unbiased is not the same as evidence that it is.
- Oversight needs to cover how systems are used: training, safeguards, accuracy, transparency, and avenues for accountability all matter.
- A response—such as a platform ban, policy change, settlement, or detection framework—is not by itself proof that the underlying harm has ended.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
