Rod2ik 🇪🇺 🇨🇵 🇪🇸 🇺🇦 🇩🇰 🇨🇦 🇬🇱<p>À peine lancé, le nouveau modèle de <a href="https://mastodon.social/tags/ChatGPT" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ChatGPT</span></a> suscite des <a href="https://mastodon.social/tags/inqui%C3%A9tudes" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>inquiétudes</span></a> de la part de plusieurs <a href="https://mastodon.social/tags/%C3%A9quipes" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>équipes</span></a> de <a href="https://mastodon.social/tags/s%C3%A9curit%C3%A9s" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>sécurités</span></a> . Des <a href="https://mastodon.social/tags/chercheurs" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>chercheurs</span></a> ont vu à plusieurs occasions que le <a href="https://mastodon.social/tags/chatbot" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>chatbot</span></a> était capable de <a href="https://mastodon.social/tags/mentir" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>mentir</span></a> intentionnellement.</p><p><a href="https://www.frandroid.com/culture-tech/intelligence-artificielle/openai-chatgpt/2432924_le-nouveau-modele-de-chatgpt-peut-mentir-a-ses-utilisateurs" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">frandroid.com/culture-tech/int</span><span class="invisible">elligence-artificielle/openai-chatgpt/2432924_le-nouveau-modele-de-chatgpt-peut-mentir-a-ses-utilisateurs</span></a></p>