“AI Detectors Don’t Work. Here’s What to Do Instead.” MIT Sloan Teaching & 

Learning Technologies, 2024, mitsloanedtech.mit.edu/ai/teach/ai-detectors-dont-work/.

MIT Sloan Teaching & Learning Technologies states that, “AI detection software is far from foolproof—in fact, it has high error rates and can lead instructors to falsely accuse students of misconduct” (MIT). This is an article published by MIT. I will use this to support my claim that AI detectors don’t work, and my action for the reader, that teachers shouldn’t be relying on these detectors entirely. This is different from other sources because they don’t usually talk about how teachers should stop using them, they only talk about how they are unreliable. I also love this source because they have so many sources at the bottom and I can use any of those to include in my writing, it makes it much easier to find good sources.

Liang, Weixin, et al. “GPT Detectors Are Biased against Non-Native English Writers.” 

ArXiv (Cornell University), 5 Apr. 2023,

https://doi.org/10.48550/arxiv.2304.02819.

This study–conducted by Weixin Liang, Mert Yuksekgonul, Yining Mao, Eric Wu, and James Zou–found that AI detectors ”consistently misclassify non-native English writing samples as AI-generated, whereas native writing samples are robustly identified as human-written" (Liang et al.). This is a university study. It proves that AI detectors are not just unreliable, but are also mistaking text from writers whose first language is not English as AI. This is more evidence that could support my claim. It supports my argument that AI detectors are unreliable and are not 100% accurate. This source is different from the OpenAI article. Instead of just sharing OpenAI’s admission of guilt, this is third-party, academic proof of why the detectors are unreliable and flawed.

OpenAI. “New AI Classifier for Indicating AI-Written Text.” Openai.com, 2023, 

openai.com/index/new-ai-classifier-for-indicating-ai-written-text/.

OpenAI states, “As of July 20, 2023, the AI classifier is no longer available due to its low rate of accuracy.” (OpenAI) This is direct evidence to support my claim that OpenAI discontinued their AI detector. I will include this with any paragraph that mentions OpenAI and use it to support my claim that AI detectors are unreliable. This is way better than any source because the source is directly supporting my claim, it is a primary source to my topic.

Robert, Amanda. “Art Generated by AI Can’t Be Copyrighted, DC Court Says.” ABA Journal, 

25 Aug. 2023, www.abajournal.com/news/article/art-generated-by-ai-cant-be-copyrighted-dc-court-says. Accessed 26 Feb. 2026.

Amanda Robert states that, "The Copyright Office and the District of Columbia agreed that only works created by human beings are eligible for copyright.” This source will help to support my claim that you cannot claim AI content as your own. I must keep in mind that this is only referring to AI art, not writing. This source will help support any part of my essay that I want to talk about art. This source is different from most only because of how direct the information was.

U.S. Copyright Office. “Copyright Registration Guidance: Works Containing Material Generated 

by Artificial Intelligence.” U.S. Copyright Office, U.S. Copyright Office, 13 Mar. 2023, 

www.copyright.gov/ai/ai_policy_guidance.pdf. Accessed 26 Feb. 2026.

As the U.S. Copyright Office explains, the key question is "whether the 'work' is basically one of human authorship, with the computer... merely being an assisting instrument." This is the official document from the U.S. Copyright Office. The point of the document is that pure AI output cannot be claimed, but a human can be the author if their own creative contribution is the main part of the work. I can use this to explain the legality behind using AI in public works. This source is much different than any other source because it’s not an argument or an opinion, it's the real legal document from the government that my entire argument is based on.