{"id":150635,"date":"2026-10-27T08:50:34","date_gmt":"2026-10-27T08:50:34","guid":{"rendered":"https:\/\/www.europesays.com\/ch\/150635\/"},"modified":"2026-10-27T08:50:34","modified_gmt":"2026-10-27T08:50:34","slug":"how-reliable-are-ai-detectors-swiss-experts-explain","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/ch\/150635\/","title":{"rendered":"How reliable are AI detectors? Swiss experts explain"},"content":{"rendered":"<p>    <img src=\"https:\/\/www.europesays.com\/ch\/wp-content\/uploads\/2026\/10\/000_C9AR2L4.jpg\" width=\"7488\" height=\"4994\" alt=\"The writer Thelyson Orelien amongst the books\" loading=\"eager\" decoding=\"sync\" fetchpriority=\"high\"\/><\/p>\n<p>                The Haitian-Canadian writer Th\u00e9lyson Or\u00e9lien has been accused of using artificial intelligence to write his novel &#8220;C\u2019\u00e9tait \u00e7a ou mourir&#8221;.            <\/p>\n<p>            Afp Or Licensors        <\/p>\n<p>        When a detector flagged a bestselling novel as AI-generated, it sparked a controversy that had serious consequences for the author. Swiss experts and artists examine how much these tools can really prove and what their verdicts mean for creative work.\n<\/p>\n<p>        This content was published on    <\/p>\n<p>        October 27, 2026 &#8211; 09:00\n<\/p>\n<p class=\"wp-block-paragraph\">Haitian-Canadian writer Th\u00e9lyson Or\u00e9lien had a promising career ahead of him. His novel C\u2019\u00e9tait \u00e7a ou mourir (It was either that or die), a critical and commercial success, was in the running for some of France\u2019s most prestigious literary prizes, including the Goncourt. That was until an anonymous account on X accused him of writing almost the entire book using artificial intelligence (AI).<\/p>\n<p class=\"wp-block-paragraph\">Or\u00e9lien denied using AI to write the novel. But to his accusers, the evidence seemed unequivocal: Pangram, one of the most popular tools for detecting AI content, had classified more than 95% of the book as AI-generated.<\/p>\n<p class=\"wp-block-paragraph\">From there, things took a turn for the worse for Or\u00e9lien. His novel was removed from the Goncourt selection, also amid revelations by the press of plagiarism in earlier works, for which he apologised. His Canadian publisher cut ties with him after an external analysis uncovered further evidence of AI use. Or\u00e9lien continued to deny having written the book with AI, while acknowledging that he had turned to it for historical research during the creative process.<\/p>\n<p class=\"wp-block-paragraph\">From the beginning, the controversy divided the literary world between those who defended the author and questioned the reliability of AI detectors, and those who criticised him.<\/p>\n<p>    <img src=\"https:\/\/www.europesays.com\/ch\/wp-content\/uploads\/2026\/10\/hendry-2x3-crop.jpg\" width=\"1000\" height=\"1482\" alt=\"Manuel Hendry\" loading=\"lazy\" decoding=\"async\" fetchpriority=\"auto\"\/><\/p>\n<p>                Manuel Flurin Hendry is a Swiss director, screenwriter and lecturer.<\/p>\n<p>            Copyright 2017 Fabian Frost        <\/p>\n<p class=\"wp-block-paragraph\">Beyond Or\u00e9lien\u2019s case, Swiss filmmaker, screenwriter and lecturer Manuel Flurin Hendry believes it\u2019s unfair to damage a writer\u2019s reputation based on a software-generated score. \u201cThe idea that you feed a text into Pangram and then socially execute a writer feels unacceptable,\u201d Hendry told Swissinfo in an interview at the AI+X Summit in early October. Hendry defends artists\u2019 freedom to choose their own tools.<\/p>\n<p class=\"wp-block-paragraph\">Other Swiss writers and artists have openly experimented with AI. Author Charles Lewinsky incorporated entire passages generated by ChatGPT into his novel T\u00e4uschend echt. Photographer Matthieu Gafsou has used image-generation tools such as DALL\u00b7E in his artistic work.<\/p>\n<p class=\"wp-block-paragraph\">These are just a few transparent examples of how AI has entered creative work in Switzerland. But deciding how to respond to undisclosed and potentially extensive use of AI is far more difficult. As is determining how much AI-generated material is acceptable in a work of art, and whether tools like Pangram can really be trusted as a definitive proof of fraud.\u00a0<\/p>\n<p>How reliable is Pangram?<\/p>\n<p>    <img src=\"https:\/\/www.europesays.com\/ch\/wp-content\/uploads\/2026\/10\/Hendrik-1-crop.jpg\" width=\"1000\" height=\"667\" alt=\"Hendrik Strobelt\" loading=\"lazy\" decoding=\"async\" fetchpriority=\"auto\"\/><\/p>\n<p>                Hendrik Strobelt is a senior researcher at IBM in the United States.            <\/p>\n<p>            Courtesy of Hendrik Strobelt        <\/p>\n<p class=\"wp-block-paragraph\">For Hendrik Strobelt, a senior researcher at IBM in the United States, the question of Pangram\u2019s reliability is complex. In 2019, Strobelt and his colleagues developed GLTR, one of the first experimental tools designed to help identify text potentially generated by language models.<\/p>\n<p class=\"wp-block-paragraph\">At the time, OpenAI had just released GPT-2, the model on which GLTR was based. AI models tended to produce rather dull texts that were very similar in style, because this type of content was the most common, Strobelt explains.<\/p>\n<p class=\"wp-block-paragraph\">Identifying whether a text might have been produced with the help of AI was therefore relatively straightforward: it was enough to look for the most likely words.<\/p>\n<p class=\"wp-block-paragraph\">Then language models became more sophisticated. Strobelt believes that, with today\u2019s advanced AI systems, it has become extremely difficult to establish clear rules to distinguish AI-generated text from human writing.<\/p>\n<p class=\"wp-block-paragraph\">\u201cNow these models are significantly larger. They can imitate styles if you prompt them accordingly,\u201d the IBM researcher says.<\/p>\n<p>    <img src=\"https:\/\/www.europesays.com\/ch\/wp-content\/uploads\/2026\/10\/GLTR-1.png\" width=\"1000\" height=\"847\" alt=\"GLTR example\" loading=\"lazy\" decoding=\"async\" fetchpriority=\"auto\"\/><\/p>\n<p>                The tool developed by Strobelt highlighted the statistically most predictable words \u2013 potentially generated by AI \u2013 in green, the less common ones in yellow, and the most surprising ones \u2013 likely of human origin \u2013 in red or purple.            <\/p>\n<p>            Courtesy of Hendrik Strobelt        <\/p>\n<p class=\"wp-block-paragraph\">Software such as Pangram uses a method known as \u201csupervised classification\u201d. It compares and classifies texts written before generative AI became widespread \u2013 which is presumed to be human-written \u2013 with \u201cmirror\u201d versions on the same topic, generated by different commercially available language models. During training the software learns to identify the differences between them.<\/p>\n<p class=\"wp-block-paragraph\">The logic is that if a system can distinguish between texts that are very similar in subject matter, its predictions will become more accurate.<\/p>\n<p class=\"wp-block-paragraph\">\u201cYou want to train on things that are very close to each other, to be sure that you get the hard tasks trained first,\u201d Strobelt explains.<\/p>\n<p class=\"wp-block-paragraph\">Pangram <a href=\"https:\/\/www.pangram.com\/\" target=\"_blank\" rel=\"nofollow noopener\">claimsExternal link<\/a> an accuracy rate of 99.98% and a false-positive rate of 0.01%. Its high performance has also been supported by <a href=\"https:\/\/bfi.uchicago.edu\/working-papers\/artificial-writing-and-automated-detection\/\" target=\"_blank\" rel=\"nofollow noopener\">independent researchExternal link<\/a>, including a <a href=\"https:\/\/link.springer.com\/article\/10.1007\/s40979-026-00226-w\" target=\"_blank\" rel=\"nofollow noopener\">studyExternal link<\/a> published this year.<\/p>\n<p class=\"wp-block-paragraph\">According to Paolo Italiani, an AI and machine-learning researcher at the Universit\u00e0 della Svizzera italiana (USI), the longer a text is, the more reliable a detector\u2019s assessment tends to be.<\/p>\n<p class=\"wp-block-paragraph\">\u201cIt\u2019s quite probable that at least a portion of [the text] is AI-generated,\u201d he says.<\/p>\n<p>The limits of AI detectors<\/p>\n<p class=\"wp-block-paragraph\">Yet it remains difficult to understand exactly which features of a text these tools rely on to make their classifications, says Lonneke van der Plas, professor of natural language processing at USI.<\/p>\n<p class=\"wp-block-paragraph\">Interpretability studies would be needed to gain more understanding, she explains. But because commercial tools such as Pangram are largely black boxes whose internal workings are not accessible to independent researchers, it is challenging to study how they reach their conclusions.<\/p>\n<p>    <img src=\"https:\/\/www.europesays.com\/ch\/wp-content\/uploads\/2026\/10\/Lonneke_Italian_cropi.jpg\" width=\"1500\" height=\"1000\" alt=\"Lonneke van der Plas and Paolo Italiani\" loading=\"lazy\" decoding=\"async\" fetchpriority=\"auto\"\/><\/p>\n<p>                AI detection tools lose accuracy when dealing with non-English texts and with content written in collaboration with AI, according to Lonneke van der Plas and Paolo Italiani of the Universit\u00e0 della Svizzera italiana.            <\/p>\n<p>            Universita\u2019 Della Svizzera Italiana        <\/p>\n<p class=\"wp-block-paragraph\">In theory, certain clues such as an overly linear argumentative structure could help determine whether a text was more likely written by AI or by a human, <a href=\"https:\/\/arxiv.org\/pdf\/2610.01353\" target=\"_blank\" rel=\"nofollow noopener\">researchExternal link<\/a> co-authored by Van der Plas highlights. But such features, she cautions, are not sufficient to establish a text\u2019s origin with certainty.<\/p>\n<p class=\"wp-block-paragraph\">Van der Plas also points out that detection tools tend to perform less reliably in languages other than English and in more specialised genres that may have been less well represented in their training data.<\/p>\n<p class=\"wp-block-paragraph\">The problem becomes even more complicated when a text is the product of collaboration between humans and AI. Some tests <a href=\"https:\/\/arxiv.org\/pdf\/2411.04032\" target=\"_blank\" rel=\"nofollow noopener\">showExternal link<\/a>, detectors\u2019 accuracy drops by around 20 percentage points in such cases. Pangram\u2019s own <a href=\"https:\/\/arxiv.org\/abs\/2607.27183\" target=\"_blank\" rel=\"nofollow noopener\">technical reportExternal link<\/a> also indicates that the software loses accuracy when faced with content that alternates between human-written and AI-generated sentences.<\/p>\n<p class=\"wp-block-paragraph\">\u201cThis software has been mostly trained on either completely human-generated or machine-generated texts. As soon as they get something that it\u2019s in between, they get into trouble,\u201d van der Plas says.<\/p>\n<p>A score is not sufficient proof<\/p>\n<p class=\"wp-block-paragraph\">From a legal perspective, results produced by tools such as Pangram do not constitute proof that an author has committed fraud.<\/p>\n<p>    <img src=\"https:\/\/www.europesays.com\/ch\/wp-content\/uploads\/2026\/10\/De_Werra-crop.png\" width=\"968\" height=\"1036\" alt=\"Jacques de Werra\" loading=\"lazy\" decoding=\"async\" fetchpriority=\"auto\"\/><\/p>\n<p>                Jacques de Werra is professor of intellectual property law at the University of Geneva.            <\/p>\n<p>            Courtesy of Jacques de Werra        <\/p>\n<p class=\"wp-block-paragraph\">\u201cA legal decision cannot simply rely on those technologies because they can be inaccurate\u201d says Jacques de Werra, professor of intellectual property law at the University of Geneva.<\/p>\n<p class=\"wp-block-paragraph\">If a case like Or\u00e9lien\u2019s were brought before a court, de Werra explains, other evidence alongside detector scores should be considered, including original manuscripts, successive drafts and even exchanges with AI systems. De Werra also points out that AI involvement in creative work does not automatically invalidate an author\u2019s copyright claims.<\/p>\n<p class=\"wp-block-paragraph\">\u201cIf a person can demonstrate that there is a sufficiently close creative link between the input, creative prompting, and the output, it could be possible to claim human creativity,\u201d he says.<\/p>\n<p class=\"wp-block-paragraph\">The key is whether the author\u2019s creative contribution is sufficient. De Werra stresses, however, that there is no fixed numerical threshold that determines how much of a text must be written by a person for a work to be considered their own.<\/p>\n<p>If a book moves you, does it matter if AI helped write it?<\/p>\n<p class=\"wp-block-paragraph\">For Manuel Flurin Hendry, however, the issue goes beyond legal protection.<\/p>\n<p class=\"wp-block-paragraph\">Hendry openly uses AI-based technologies, some of which he has developed himself, for various aspects of his work, from analysing screenplays to assisting his academic research.<\/p>\n<p class=\"wp-block-paragraph\">Technology, he explains, opens new channels of inspiration, sometimes suggesting approaches he had not previously considered. It also frees up more time for the creative process.<\/p>\n<p class=\"wp-block-paragraph\">\u201cTrue human writing is thinking in action. It requires friction and to enter a state where one can just concentrate on their flow of thoughts,\u201d he says.<\/p>\n<p class=\"wp-block-paragraph\">And he poses a provocative question: why should it matter how much AI contributed to a work? \u201cIf it moves you. If you find it interesting. If it makes you think, what is the argument?\u201d he asks.<\/p>\n<p class=\"wp-block-paragraph\">What\u2019s important, Hendry argues, is learning which tasks to delegate to AI and which are better done personally. Rather than rejecting technology, artists should be transparent about how they use it.<\/p>\n<p class=\"wp-block-paragraph\">\u201cIf I pretend this is mine, but it\u2019s totally AI generated, then it\u2019s fake,\u201d Hendry says.<\/p>\n<p class=\"wp-block-paragraph\">What is at stake is not only an artist\u2019s reputation, but above all the relationship of trust with their audience.<\/p>\n<p class=\"wp-block-paragraph\">Edited by Gabe Bullard\/VdV\u00a0<\/p>\n","protected":false},"excerpt":{"rendered":"The Haitian-Canadian writer Th\u00e9lyson Or\u00e9lien has been accused of using artificial intelligence to write his novel &#8220;C\u2019\u00e9tait \u00e7a&hellip;\n","protected":false},"author":2,"featured_media":150636,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":"","_share_on_mastodon":"0"},"categories":[4],"tags":[101,9334,334,17,9336],"class_list":["post-150635","post","type-post","status-publish","format-standard","has-post-thumbnail","category-switzerland","tag-article","tag-beat-swiss-ai","tag-production-type-adaptation","tag-switzerland","tag-technology-general"],"share_on_mastodon":{"url":"https:\/\/pubeurope.com\/@ch\/117512019842433633","error":""},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/posts\/150635","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/comments?post=150635"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/posts\/150635\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/media\/150636"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/media?parent=150635"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/categories?post=150635"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/tags?post=150635"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}