\u00c0 l'occasion du Festival litt\u00e9raire et culturel de Babell \u00e0 Porto, au Portugal, l'\u00e9minente autrice Margaret Atwood, surtout connue pour ses livres The Handmaid’s Tale<\/em> et L'Assassin aveugle<\/em>, a particip\u00e9 \u00e0 une discussion franche sur l'IA et a exprim\u00e9 certaines r\u00e9serves.<\/p>\n AI’s omnipresence in modern society almost guarantees it to be a hot topic in forums and discussions. This festival was no exception, with the esteemed Canadian author sharing anecdotes and personal experiences with AI technology. Most notably, she shared her less than satisfactory encounter with Anthropic’s AI chatbot, Claude.<\/p>\n Atwood’s introduction to AI chatbots, it seems, was less than stellar. According to the acclaimed writer, she had turned to Claude, like one would consult a convenient oracle, for some specific information on the British detective series Le p\u00e8re Brown<\/em>. La r\u00e9ponse qu'elle a re\u00e7ue l'a toutefois laiss\u00e9e sur sa faim et a \u00e9veill\u00e9 ses soup\u00e7ons.<\/p>\n “Claude gave me the wrong answer, or it lied. Of course, it didn’t know it was lying because it’s not a human being; it’s a large language model\u2026 It had skimmed a \u2026<\/p>\n<\/blockquote>\n Cette erreur de Claude met en \u00e9vidence une faille inh\u00e9rente \u00e0 la technologie de l'IA. Les robots, d\u00e9pourvus d'intuition et d'exp\u00e9rience humaines, peuvent finir par d\u00e9former les faits, ce qui conduit \u00e0 la d\u00e9sinformation, un ph\u00e9nom\u00e8ne particuli\u00e8rement n\u00e9faste \u00e0 l'\u00e8re de la surabondance d'informations.<\/p>\n Despite being framed as advanced and accurate, AI can stumble, as depicted in Atwood’s anecdote. This fallibility of AI has a resonance with Atwood’s extensive exploration of dystopia and the dangers of uncontrolled technological advancement in her works.<\/p>\n Does this mean we should pull the plug on AI development? Surely not. But much like Margaret Atwood’s tales, her experiences shed a cautionary light on our relationship with technology and demand a more holistic approach to understanding and integrating AI into our daily lives.<\/p>\nQuand les chatbots bas\u00e9s sur l'IA se trompent<\/h2>\n
\n