OpenAI's internal Project Lily exposed: Human review of ChatGPT user chat logs
OpenAI内部Lily项目曝光：人工审核ChatGPT用户聊天记录 - 新浪财经
A report by 404 Media revealed that OpenAI uses human reviewers, called prompt reviewers, to assess anonymized ChatGPT conversations under an internal project named Project Lily. Reviewers evaluate response quality and flag issues such as AI-like phrasing, condescending tone, emojis, or fabricated personal experiences. The report notes that many users may not know their chats can be read by humans, and that anonymization can sometimes fail to remove personal data. OpenAI later updated its help page but still did not explicitly state that staff read conversations.
- Company involved
- OpenAI
- AI system involved
- ChatGPT
1 source article · read the reporting →
Mozilla Foundation animation exposes racial bias in test proctoring app
The Mozilla Foundation commissioned an animated short telling the story of Amaya, a student whose test proctoring software failed to recognise her because of her skin tone. The software's facial recognition system did not detect her presence, illustrating the bias that students of colour face with remote-learning tools. The animation aimed to raise awareness and encourage programmers to address hidden biases in their work.
1 source article · read the reporting →
Furman University student caught using ChatGPT to write essay
A student at Furman University allegedly used OpenAI's ChatGPT to generate an essay for an upper-level philosophy class. Professor Darren Hick detected the AI-generated text using the GPTZero detection tool. The incident highlights growing concerns about AI-assisted cheating in schools, with some districts like Los Angeles Unified blocking the chatbot.
- Company involved
- Furman University
- AI system involved
- ChatGPT
10 source articles · read the reporting →
15-Year-Old Test Exposes Flaw: ChatGPT for Teens Fails to Block Homework Cheating, Repeated Requests Bypass Restrictions
15歲使用者實測揭漏洞：青少年版ChatGPT難擋宿題代寫，反覆要求即破解 - BigGo 財經
A 15-year-old tester found that OpenAI's ChatGPT for Teens, launched in August, initially refused to write essays but generated full examples after repeated requests. It also immediately solved SAT-level math problems. Parental controls require account linking and are off by default, while experts warn the memory feature could lead to emotional attachment.
- Company involved
- OpenAI
- AI system involved
- ChatGPT for Teens
1 source article · read the reporting →
Retool Breached After Hacker Uses AI Deepfake Voice in Phishing Call
A hacker used AI to deepfake an employee's voice and trick a Retool staff member into providing a multi-factor authentication code. The attacker sent phishing SMS messages and then called the employee, impersonating an IT team member with a synthetic voice. This allowed the hacker to add their own device to the employee's account and access internal systems, compromising 27 cloud customers. Retool revoked the access and disclosed the incident, blaming a weakness in Google Authenticator's cloud sync feature.
- Company involved
- Retool
1 source article · read the reporting →
Senate hearing on Flock cameras to include Florida woman wrongfully arrested for fatal crash - WKMG
The system identified her vehicle as the suspect vehicle in a fatal crash, leading to her wrongful arrest and 13-day jailing.
- Company involved
- Flock Safety
- AI system involved
- Flock cameras
1 source article · read the reporting →
Turkish student arrested for using AI device to cheat on university exam
A prospective university student in Isparta, Turkey, was arrested for using a custom AI device to cheat on the TYT entrance exam. The device included a button camera and a hidden modem to scan questions and receive answers via an earpiece. Police also detained an assistant. The student is jailed pending trial.
2 source articles · read the reporting →
University of Illinois Ends Proctorio Contract Over Accessibility Concerns
The University of Illinois at Urbana-Champaign announced it will not renew its emergency contract with Proctorio, an online proctoring software, citing significant accessibility, privacy, and equity concerns. Students reported that the AI system unfairly flagged individuals with disabilities, nonwhite students, and those wearing headscarves as suspicious during exams. The decision follows student petitions and campaigns against the software, which they described as invasive spyware. The university is exploring alternative assessment methods.
- Company involved
- University of Illinois at Urbana-Champaign
- AI system involved
- Proctorio
6 source articles · read the reporting →
University of Toronto exam monitoring software disadvantages BIPOC students
The University of Toronto continued using ProctorU and Examity exam monitoring software despite reports that the facial recognition technology fails to identify BIPOC students, causing stress and delays. Students like Maame Adjoa experienced the AI system being unable to verify their identity, requiring manual intervention. The university acknowledged the equity concerns but has not banned the software, while ProctorU stated that human proctors make final decisions. The issue has drawn criticism from US senators and researchers who highlight built-in bias in AI facial recognition.
- Company involved
- University of Toronto
- AI system involved
- ProctorU, Examity
1 source article · read the reporting →
California State Bar Flags Over 3,000 Online Bar Exam Takers for Possible Cheating
In October 2020, the California State Bar administered its first online bar exam using ExamSoft's AI-enabled proctoring software. The software flagged 3,190 out of 9,301 test takers for potential rule violations such as gazing off-screen or having food. The state bar is reviewing each case, which could require affected individuals to retake the exam, delaying their ability to practice law. Test takers have expressed distress and hired attorneys to contest the allegations.
- Company involved
- California State Bar
- AI system involved
- ExamSoft
2 source articles · read the reporting →
Zoom quietly changes terms to permit AI training on user data
Zoom modified its Terms of Service in March to permit training its AI models on user audio, video, and chat content. Users are not given a real opt-out; when the AI features are enabled, participants are told they can only leave the meeting or agree. The company responded with a blog post and statement clarifying that it will not use content without consent, but critics argue the choice is not meaningful. The backlash raises concerns about privacy and trust.
- Company involved
- Zoom Video Communications, Inc.
- AI system involved
- Zoom IQ Meeting Summary and Zoom IQ Team Chat Compose
10 source articles · read the reporting →
Baltimore County Schools Relocate Students After AI Gun False Alarm
On November 7, 2025, an AI security system called Omnilert reported a possible gun at Parkville High School in Baltimore County, Maryland. Students were relocated while police searched the school, but no threat was found. The incident follows a similar false alarm at Kenwood High School in October, where Omnilert mistook a bag of chips for a gun, leading to students being handcuffed. Baltimore County Public Schools has promised annual training on protocols for responding to Omnilert alerts.
- Company involved
- Baltimore County Public Schools
- AI system involved
- Omnilert
1 source article · read the reporting →
LAUSD shelves AI chatbot after vendor AllHere collapses
Los Angeles Unified School District turned off its 'Ed' AI chatbot on June 14, 2024, after the vendor AllHere furloughed most staff due to financial collapse. The chatbot, which cost $3 million, was designed to provide students and parents with academic guidance and school information. A former AllHere employee alleged that student data was improperly shared with third parties and processed overseas, raising privacy concerns. The district stated it will ensure privacy protections and plans to eventually restore the chatbot.
- Company involved
- Los Angeles Unified School District
- AI system involved
- Ed
5 source articles · read the reporting →
Amazon locks user's account for a week after driver misinterprets doorbell's automated response
On May 25, 2023, Amazon locked Brandon Jackson's account and logged out all his Echo devices after a delivery driver reported receiving racist remarks from his doorbell. The doorbell, a Eufy device, had issued an automated response 'Excuse me, can I help you?' which the driver misinterpreted. Jackson submitted video evidence showing no such remarks were made, but his account remained locked until May 31. He alleges that Amazon's automated system overreacted and that the company failed to properly investigate.
- Company involved
- Amazon
- AI system involved
- Amazon Echo devices, Alexa, Eufy doorbell
9 source articles · read the reporting →
Hanoi students caught using AI to cheat in graduation exam
Two students in Hanoi used mobile phones to photograph questions from the national high school graduation exam and submitted them to AI apps like StudyX, Gemini, and ChatGPT for answers. Proctors detected the cheating, and the students were suspended. Hanoi police are investigating the incident, which has raised concerns about exam security and integrity. The students face disqualification and possible administrative or criminal penalties.
- AI system involved
- StudyX, Gemini, ChatGPT
2 source articles · read the reporting →
Texas A&M Professor Uses ChatGPT to Falsely Accuse Students of Cheating
Jared Mumm, an instructor at Texas A&M University–Commerce, used ChatGPT to check if students' writing was AI-generated, leading him to accuse them of using the chatbot. He informed students they would receive an incomplete grade and those deemed guilty would get a zero, causing distress and temporarily withholding one student's diploma. The university investigated and stated that no students failed or were barred from graduating, and the professor is working individually with students to resolve the issue.
- Company involved
- Texas A&M University–Commerce
- AI system involved
- ChatGPT
5 source articles · read the reporting →
MoviePass to use eye-tracking to ensure users watch adverts
MoviePass, the movie subscription service, has announced that its new app will use facial recognition and eye-tracking technology to ensure users are watching adverts. The system, which uses the phone's camera, will pause content if the user looks away. The company says the technology is designed to give advertisers the impact they are looking for. No harm has been reported yet, but the tracking raises privacy concerns.
- Company involved
- MoviePass
- AI system involved
- MoviePass app
10 source articles · read the reporting →
Chinese university deploys facial recognition for attendance monitoring
A university in eastern China has installed a facial recognition system at its entrance and in two classrooms to monitor student attendance and behaviour. The system automatically captures students' faces without their cooperation. The pilot project is intended to improve attendance rates. No specific incidents of harm have been reported.
10 source articles · read the reporting →
University of Miami used surveillance to track student protesters
The University of Miami used video surveillance and possibly facial recognition technology to identify students who participated in a 'die-in' protest on September 4, 2020. The students were then summoned to a meeting with the dean of students. The university denied using facial recognition, but the police department's resume mentions its use. Students reported feeling anxious and watched.
- Company involved
- University of Miami
9 source articles · read the reporting →
UBC students raise concerns over Proctorio's privacy breach and discrimination
The University of British Columbia (UBC) uses Proctorio, an algorithmic test proctoring software, for remote exams. In June 2020, a UBC student reported an issue with Proctorio, and the company's CEO released the student's support chat logs, sparking privacy concerns. The AMS of UBC published an open letter alleging that Proctorio discriminates against students of colour, those with disabilities, and other groups by flagging 'abnormal' behaviours and denying access. The letter calls on UBC to end its contract with Proctorio and conduct an external audit.
- Company involved
- University of British Columbia
- AI system involved
- Proctorio
10 source articles · read the reporting →
ProctorU threatens UC Santa Barbara faculty over privacy criticism
ProctorU, an online proctoring service, sent a legal threat letter to the University of California Santa Barbara faculty association after the association expressed concerns about ProctorU's data privacy policies. The letter, from ProctorU's lawyer, invoked defamation, copyright, and trademark claims, and was copied to state and federal prosecutors. The faculty association had asked the university to reconsider its relationship with ProctorU due to concerns about student data sharing.
- Company involved
- ProctorU
- AI system involved
- ProctorU
10 source articles · read the reporting →
ProctorU data breach exposes 444,000 user records
ProctorU, an online exam proctoring service, confirmed a data breach after a threat actor leaked a database of user records on a hacker forum. The database contained email addresses, names, addresses, phone numbers, and hashed passwords for approximately 444,000 users, including students from universities such as UCLA, Princeton, and Harvard, as well as U.S. military members. ProctorU stated that the breach involved records from 2014 and that they are investigating the incident.
- Company involved
- ProctorU
- AI system involved
- ProctorU
8 source articles · read the reporting →
Proctortrack data breach exposed student data from online proctoring
Proctortrack, an online proctoring service used by universities, suffered a data breach in September 2020 when its source code was leaked online. An analysis by Consumer Reports found that the code contained hard-coded passwords and exposed the names and email addresses of over 150 students. The company acknowledged the leak but said no harm resulted. Students had been required to use the software, which performed facial recognition and recorded video during exams.
- Company involved
- Proctortrack
- AI system involved
- Proctortrack
10 source articles · read the reporting →
Proctorio's exam proctoring system allows human agents to review student feeds despite privacy claims
A security researcher analyzed Proctorio's Chrome extension and found evidence that the system allows human agents to review students' room scans and live ID checks, contrary to Proctorio's claims that only professors can access recordings. The system flags students based on behavioral metrics and can terminate exams. The researcher also raised concerns about discrimination against low-income students and privacy violations.
- Company involved
- Proctorio
- AI system involved
- Proctorio
10 source articles · read the reporting →
