New Tool Detects ChatGPT-Generated Academic Text With 99% Accuracy — Heather Desaire, a Univers |
Scientists have developed an AI-detection tool with 99% accuracy that identifies scientific text generated by AI, specifically targeting the likes of ChatGPT. The detection tool, uniquely designed using a smaller dataset and human insight to distinguish AI from human writing, narrows its focus to the specific scientific writing found in peer-reviewed journals, offering more accuracy than general-purpose detectors.
She said the success of her detection method depends on narrowing the scope of writing under scrutiny to scientific writing of the kind found commonly in peer-reviewed journals. This improves accuracy over existing AI-detection tools, like the RoBERTa detector, which aim to detect AI in more general writing.
“When you start to think about ‘AI plagiarism,’ 90% accurate isn’t good enough,” Desaire said. “You can’t go around accusing people of surreptitiously using AI and be frequently wrong in those accusations — accuracy is critical. But to get accuracy, the trade-off is most often generalizability.” “We used a much smaller dataset and much more human intervention to identify the key differences for our detector to focus on,” Desaire said. “To be exact, we built our strategy using just 64 human-written documents and 128 AI documents as our training data. This is maybe 100,000 times smaller than the size of data sets used to train other detectors. People often gloss over numbers. But 100,000 times — that’s the difference between the cost of a cup of coffee and a house.
In another important aspect, Desaire and her group flipped the script on methods used by previous teams building AI-detection methods. Desaire has made her team’s AI-detecting code fully accessible to researchers interested in building off it. She hopes others will realize that AI and AI detection are within reach of people who might not consider themselves computer programmers now.
France Dernières Nouvelles, France Actualités
Similar News:Vous pouvez également lire des articles d'actualité similaires à celui-ci que nous avons collectés auprès d'autres sources d'information.
Brooklyn 99 Secretly Confirmed A Wild Rosa Diaz Claim Was Totally TrueA secret Brooklyn Nine-Nine detail confirms a wild Rosa Diaz claim to be true and undoubtedly explains the tendencies of her mysterious character.
Lire la suite »
This new tool helps teenagers think about careers that will actually last | CNN BusinessLike many adults, a lot of teenagers don’t know what they want to do when they grow up.
Lire la suite »
Stable Diffusion Releases New AI Outpainting Tool Called UncropUncrop will change the aspect ratio of any image by very simply dragging the crop perimeter any way you like.
Lire la suite »
Lawyers blame ChatGPT for tricking them into citing bogus case lawTwo attorneys are facing possible punishment over a filing in a lawsuit against an airline that included references to past court cases that they thought were real, but were actually invented by the artificial intelligence-powered chatbot.
Lire la suite »
Inside Amazon's rush to respond to ChatGPTInside Amazon's rush to respond to ChatGPT. 'Amazon must be really scared about being late on all of this.'
Lire la suite »
Meta's Open-Source 'MusicGen' AI Is Like ChatGPT for TunesMeta has released another open-source AI model trained on hundreds of thousands of music tracks online.
Lire la suite »