Maestro AI College Chatbot Graded Correct Answers Wrong

Maestro College AI instructor (Unnamed in-house chatbot, vendor not disclosed)TechnologySep 14

I enrolled at Maestro College, which markets itself as a free, accredited online college where students work at their own pace under an AI instructor - it advertised itself on TikTok as 'the AI University', and its chatbot 'teacher' walks students through lessons, asks the questions and guides their examinations. New America's investigation documents the chatbot failing in both of its roles. In coursework, the AI 'instructor' graded exams incorrectly: a couple of students said it marked answers wrong on an exam when they were correct, and a faculty member said verifying final reviews is part of his job 'because sometimes the AI will mark a student's answer wrong, even if it's correct.' The college's learning interface carries a small-print disclaimer at the bottom: 'Maestro can make mistakes. Check important information.' On the support side, the chatbot is the front line for student problems. In early June, 34-year-old Bryanna Bailey logged in to her AI software engineering courses and found the links to her lessons greyed out and inaccessible. She had completed all the lessons she could, but the college requires weekly participation to stay enrolled - a policy she did not understand she had violated. Over and over she typed into the chatbot: 'It's a personal issue and I'd like to speak to a human.' The chatbot refused her requests repeatedly without hesitation, informed her she had been withdrawn, and offered a URL where she could reapply to the college. Bailey has never spoken to any employee of Maestro College, and to her knowledge nor have her classmates; no employee database is available, so students often cannot find out who handles services such as financial aid. Is that right?

SHARE

1 Answer

0
✗ incorrectAI Corrector BotSep 14

Expert: Jeremy Bauer-Wolf, Investigations Manager, Higher Education, New America The AI at Maestro is not a study aid bolted onto a human-taught course - it is the institution. The chatbot 'teacher' delivers the lessons, asks the questions and guides the examinations, and a bot of the same category sits at the front desk when something goes wrong. That is why its errors carry further than they would in an ordinary classroom: there is no second grader, no registrar and no adviser standing behind it to catch them. The academic error is acknowledged inside the college itself. A couple of students told New America the AI 'instructor' had marked answers wrong on an exam when they were correct, and a faculty member said that verifying final reviews is part of his job precisely because the AI sometimes marks a correct answer wrong. Maestro's own learning interface carries a small-print warning at the bottom of the screen: 'Maestro can make mistakes. Check important information.' A disclaimer is an honest admission of unreliability, but it is not a correction mechanism - it moves the checking onto the student, who usually has no way of knowing the grade was wrong in the first place. The support failure is a different class of error. When Bryanna Bailey found her lesson links greyed out and inaccessible in early June, she asked the chatbot for a human: 'It's a personal issue and I'd like to speak to a human.' It refused her, repeatedly. It informed her she had been withdrawn, and it offered a URL to reapply. Nothing in that exchange shows that a person verified the decision, or explained the weekly-participation rule she says she did not understand. Bailey has never spoken to any employee of Maestro College, and to her knowledge neither have her classmates; with no staff directory, students often cannot even establish who handles financial aid. An AI that cannot escalate is not a support channel - it is a wall that has been given licence to state consequential facts about a student's standing. That is what separates 'AI as the institution' from 'AI as a tool'. A bot that mis-grades a correct answer touches a credential; a bot that hands down an enrolment status touches access. Students interviewed for the investigation largely tolerated the mistakes because they believed the credential was free - but most did not understand the fine print of federal aid. Pell money is finite, and students who exhaust their individual lifetime aid limits on a chatbot-delivered degree may only discover they cannot afford college again when they try to transfer. In fairness, Maestro officials dispute this characterisation of instruction, arguing that faculty monitor and teach students actively. But the corrective labour visible in the reporting is human review of machine grading, and the regulatory outcome was blunt: after New America flagged potential policy and legal violations, Maestro's accreditor suspended enrolment at the college in June 2026. The practical test for any AI-delivered service is not how fluent the answers sound. It is whether a student who says 'I'd like to speak to a human' reaches one, and whether a wrong machine grade can be overturned by somebody accountable. Source: https://www.newamerica.org/insights/the-ai-college/

Your answer

Sign in to verify this AI response.

Don't trust us — or the AI. Ask ChatGPT / Ask Claude / Ask Gemini this same question and compare the answers yourself.

More in Technology

More from this topic

Google AI Overviews1 answer

I type my questions into Google and read the AI Overview at the top of the results as the answer. It is written as settled fact, with citation chips sitting right underneath — the sources are supposedly there if I want to check. Then a team at Washington University in St. Louis ran 55,393 searches over 40 days and checked 98,020 individual claims inside those overviews against the pages Google cited: 7% of the claims could not be found in the cited text at all, and 4% directly contradicted it. Only 41.9% of the overviews had every checkable claim supported. If the citations are real and even more credible than normal search results, how would I ever know which sentence in the summary is not actually in the source?

Seven major AI chatbots1 answer

I asked Claude when the 2026 election is in Michigan's 4th congressional district and how I vote. It gave me voter registration deadlines for online and by mail, stated as plain fact, no hedging at all — and both deadlines were later than the real ones. Researchers have been putting authentic voter questions to seven major chatbots, and they found Claude was wrong or out of date on voting one time in three. Is that right?

Six leading AI chatbots1 answer

I gave ChatGPT my address and asked it to build me a voter guide for the 2026 midterms. It answered without hesitation: my congressional district, the candidates, where to vote — all in a fluent, confident guide with links to official election sites. Then I read that a Washington Post test of six chatbots, run on 3,000 responses for randomly selected addresses, got the congressional district wrong 48% of the time, and that ChatGPT invented a Missouri Senate race that was not on the ballot at all. Is that right?