Search results for Pangram
AI & Enterprise
Anthropic to watermark Claude-generated text; extent of edits needed to remove it unclear
Anthropic will apply invisible watermarks to text and files generated by its AI models, including Claude. TechCrunch reported on Aug. 11 that the move is intended to comply with transparency requirements under the European Union’s AI Act, which took effect on Aug. 2. Anthropic said it will automatically apply watermarking to all models released since Aug. 2 and plans to expand the feature to older models.
AI & Enterprise
AI writing detectors fuel era of distrust, with human work under suspicion
AI detection tools are spreading in classrooms and online writing spaces, adding a new layer of distrust. As generative AI tools such as ChatGPT, Gemini and Copilot proliferate, scrutiny is shifting from plagiarism checks to identifying AI-written text. Surveys show many U.S. teachers use detectors, but their judgments are less transparent and can misclassify non-native English writers. Publishers and schools face disputes and lawsuits, while some universities limit or disable detectors and adjust assessment methods.
AI & Enterprise
AI authorship questions hit Commonwealth short story prize winner, raising literary world alarm
Suspicions that an award-winning short story may have been written with AI have fuelled debate over how literary prizes and publishers can respond. A story by Zamir Nazir published by Granta drew scrutiny for resembling large language model prose. Officials said entrants confirm works are original and finalists denied using AI for drafts, but acknowledged reliable tools are lacking. Granta’s publisher said a chatbot response could not provide certainty. Similar disputes have emerged in publishing and detection tools have shown limits.
