{"id":16510,"date":"2025-08-26T09:37:24","date_gmt":"2025-08-26T08:37:24","guid":{"rendered":"https:\/\/dataethics.eu\/the-ai-hype-in-a-new-guise-agentic-ai\/"},"modified":"2025-08-26T13:50:43","modified_gmt":"2025-08-26T12:50:43","slug":"the-ai-hype-in-a-new-guise-agentic-ai","status":"publish","type":"post","link":"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/","title":{"rendered":"AI hype i nye kl\u00e6der: AI agenter"},"content":{"rendered":"\n

Den generative AI-hype bobbel er ved at briste, s\u00e5 hvorfor ikke genindf\u00f8re en gammel id\u00e9 om AI agenter, s\u00e5 vi kan blive ved med at \u00f8ge AI-investeringerne? AI agenter er lige s\u00e5 gammel som AI’s historie. Men i dag er l\u00f8fterne om dens potentiale st\u00f8rre end nogensinde.<\/strong><\/p>\n\n\n\n

De sidste par \u00e5r har l\u00f8fterne om generativ AI v\u00e6ret ekstreme: Vi kan automatisere millioner af jobs og spare en masse penge p\u00e5 menneskelige medarbejdere. Vi vil se en produktivitetsstigning i alle brancher. Medievirksomheder vil lide, fordi AI kan udf\u00f8re det meste af journalisternes arbejde. Chatbots kan overtage kundeservice, l\u00f8se de fleste af vores problemer, og nu f\u00e5r vi at vide, at AI endda kan udf\u00f8re en ph.d.’s arbejde. Det kan man da kun elske? Det eneste du skal g\u00f8re er at prompte bedre. <\/p>\n\n\n\n

Den generative AI-guldfeber har varet siden 2022, hvor ChatGPT blev lanceret. Langsomt er vi begyndt at at indse, at disse l\u00f8fter er mere science fiction end videnskab. Efter alle disse \u00e5rs l\u00f8fter venter vi stadig p\u00e5 reelle effektivitetsgevinster,<\/a> modellerne hallucinerer stadig by design, kan ikke klare helt simpel matematik og beg\u00e5r fejl i de simpleste opgaver, s\u00e5som at placere lande p\u00e5 et landkort<\/a>.<\/p>\n\n\n\n

S\u00e5 hvad skal AI-virksomhederne g\u00f8re, n\u00e5r de har brug for AI-investeringer for at d\u00e6kke de stigende omkostninger ved GenAI? De har tilsyneladende genindf\u00f8rt ideen om den autonome agent \u2013 det store h\u00e5b, der siden teknologiens begyndelse har hjems\u00f8gt feltet \u2013 fra Turings t\u00e6nkende maskiner og McCarthys r\u00e5dgivere til det mere popul\u00e6re udtryk som elektroniske hjerner, der blev brugt under den f\u00f8rste AI-hype i 1950’erne. Lad os kalde det AI agenter og endnu en gang love, at denne gang, ja, denne gang helt sikkert, vil disse maskiner blive en del af arbejdsstyrken, erstatte videnarbejde og blive den st\u00f8rste leverand\u00f8r af digital arbejdskraft, if\u00f8lge TechChrunch<\/a>.<\/p>\n\n\n\n

AI-agenter er softwareprodukter, der er designet til at udf\u00f8re opgaver best\u00e5ende af flere dele p\u00e5 vegne af deres menneskelige opgavestillere, if\u00f8lge Futurism<\/a>.<\/p>\n\n\n\n

Gartner \u2013 et af de konsulentfirmaer, der er plejer at hype nye teknologier \u2013 skrev i en rapport fra juni <\/a>2025, at “mindst 15 % af de daglige arbejdsbeslutninger vil blive truffet autonomt gennem AI agenter inden 2028, hvilket er en stigning fra 0 % i 2024. Derudover vil 33 % af virksomhedens softwareapplikationer omfatte agentisk AI i 2028, hvilket er en stigning fra mindre end 1 % i 2024.”<\/p>\n\n\n\n

Det lyder ikke umiddelbart hypet, og Gartners rapport angiver da ogs\u00e5, at over 40 % af de agentiske AI-projekter vil blive aflyst inden udgangen af 2027 p\u00e5 grund af stigende omkostninger, uklar forretningsv\u00e6rdi eller utilstr\u00e6kkelig risikokontrol.<\/p>\n\n\n\n

Selvf\u00f8lgelig lever AI-agenter ikke op til l\u00f8fterne p\u00e5 nuv\u00e6rende tidspunkt. If\u00f8lge forskere ved Carnegie Mellon University, der udgav en artikel i juni 2025,<\/a> fejlede den bedst pr\u00e6sterende AI-agent, Googles Gemini 2.5 Pro, i at udf\u00f8re virkelige kontoropgaver<\/a> i 70 procent af tilf\u00e6ldene. Mens OpenAI’s GPT-4o havde en fejlprocent p\u00e5 91,4 procent, havde Meta’s Llama-3.1-405b en fejlprocent p\u00e5 92,6 procent, og Amazon’s Nova-Pro-v1 fejlede i hele 98,3 procent af sine kontoropgaver.<\/p>\n\n\n\n

AI agenter er et mareridt for privatlivets fred<\/strong><\/p>\n\n\n\n

Udfordringerne ved autonome AI-systemer \u2013 tidligere tiders \u00bbelektroniske hjerner\u00ab, nutidens \u00bbagentiske AI\u00ab \u2013 er mange og veldokumenterede. Det er ogs\u00e5 derfor, vi er n\u00e5et frem til en mere nuanceret tilgang til AI p\u00e5 det internationale omr\u00e5de for AI-regulering. I dag har vi heldigvis EU’s AI-lov, der automatisk fastl\u00e6gger en ikke-agentisk AI-vej fremad. Som det fremg\u00e5r af lovens f\u00f8rste krav: \u00bbMennesker skal bevare kontrollen over systemet og v\u00e6re i stand til at gribe ind n\u00e5r som helst for at forhindre eller minimere skader\u00ab.<\/p>\n\n\n\n

For at give et mere tydeligt eksempel p\u00e5 udfordringerne kan vi se p\u00e5 retten til privatlivets fred. AI agenter overg\u00e5r naturligvis alle de big data-teknologier, der er introduceret i l\u00f8bet af de sidste 20 \u00e5r, n\u00e5r det g\u00e6lder udfordringer i forbindelse med privatlivets fred. Mens generativ AI har suget dine data til sig i stor stil, har AI agenter brug for endnu flere. Jo mere du \u00f8nsker, at de skal betjene dig, des flere data har de brug for. Hvis du f.eks. \u00f8nsker at lave en aftale med din fysioterapeut og have din AI-agent til at g\u00f8re det uden at stille yderligere sp\u00f8rgsm\u00e5l, skal den have adgang til dit kreditkort eller din bankkonto, din kalender og din e-mail for fx at kunne minde dig om aftaler eller forberede dig p\u00e5 dem.<\/p>\n\n\n\n

Tyskland og Frankrig siger forel\u00f8big nej til AI agenter <\/a>p\u00e5 grund af de mange risici.<\/p>\n\n\n\n

“The thing that is the new buzzword, agentic AI, there is a real danger. We are giving too much control to these systems that will need access to data. It can book a ticket for a concert, look into your calendar, schedule it, and message all your friends about it. Put your brain in a drawer. The data will be sent to a cloud service, and there is a profound issue around security and privacy. It will undermine your privacy.”<\/em><\/p>\n\n\n\n

Ville du stole p\u00e5, at et kommercielt, ikke-gennemsigtigt firma handlede p\u00e5 dine vegne og give dem kontrol over din kalender, dit kreditkort og dine venners kontaktoplysninger? Ja, nogle ville m\u00e5ske g\u00f8re det uden at vide det, men de fleste ville nok sige nej tak, hvis de virkelig var informeret om de mulige konsekvenser.<\/p>\n\n\n\n

If\u00f8lge The Financial Times (AI-agenter: fra co-pilot til autopilot 7.5.25) er der en kl\u00f8ft mellem hype og virkelighed. Forfatteren af bogen Agentic Artificial Intelligence, Pascal Bornet, forklarede i avisen, at autonome biler i \u00f8jeblikket fungerer p\u00e5 niveau to til fire (ud af seks niveauer), og at AI-agenter befinder sig p\u00e5 et lignende niveau. De fleste fungerer p\u00e5 niveau to eller tre, mens nogle specialiserede systemer fungerer p\u00e5 niveau fire. Men niveau fem, hvor agenter fuldt ud forst\u00e5r, planl\u00e6gger og udf\u00f8rer komplicerede missioner med minimal menneskelig indblanding, er stadig teoretisk, sagde han.<\/p>\n\n\n\n

De mange l\u00f8fter om guld, generativ AI og AI agenter kommer oftest fra dem, der har \u00f8konomiske interesser i teknologien for f.eks. at f\u00e5 investeringerne til at forts\u00e6tte.<\/p>\n\n\n\n

F\u00f8lg ikke den kommercielle hype. F\u00f8lg videnskaben.<\/strong><\/p>\n\n\n\n

Der er en alternativ vej frem, og det er inden for det videnskabelige felt AI, som heldigvis er kommet videre fra 1950’ernes begejstring, som er smerteligt bevidst om 1970’ernes AI-vinter og m\u00e5ske ogs\u00e5 ser en ny AI-vinter i 2020’erne lige om hj\u00f8rnet (AI-vinter er et udtryk, der bruges om en periode med skuffelse og budgetnedsk\u00e6ringer efter perioder med AI-hype og brudte l\u00f8fter). Den alternative AI-vej er ogs\u00e5 en sp\u00e6ndende id\u00e9 om AI’s potentiale, men den er baseret p\u00e5 \u00e5rtiers videnskabeligt arbejde og afpr\u00f8vning af, hvad der virker, og hvad der ikke virker, omhyggelig unders\u00f8gelse af sociale og etiske konsekvenser, politiske udviklinger og internationale aftaler \u2013 s\u00e5som det arbejde, vi udf\u00f8rte i EU’s ekspertgruppe om AI, i EU generelt, UNESCO, OECD og FN om p\u00e5lidelig AI med mennesker i kontrol. Den bygger p\u00e5 den f\u00e6lles forst\u00e5else, vi n\u00e5ede frem til, om den semi-autonome rolle, vi har brug for og \u00f8nsker, at AI-teknologien skal spille i vores menneskelige samfund.<\/p>\n\n\n\n

Som Yoshua Bengio – en af verdens fremmeste AI forskere – skrev for nylig: <\/p>\n\n\n\n

\u201cThe leading AI companies are increasingly focused on building generalist AI agents – systems that can autonomously plan, act, and pursue goals across almost all tasks that humans can perform. Despite how useful these systems might be, <\/em><\/a>unchecked AI agency poses significant risks to public safety and security, ranging from misuse by malicious actors to a potentially irreversible loss of human control.”<\/em><\/p>\n\n\n\n

De foresl\u00e5r i stedet at fokusere p\u00e5 ikke-agentiske AI-systemer, der er p\u00e5lidelige og sikre i deres design. Vi kunne kalde det videnskabelig AI, som er et system designet til at forklare verden ud fra observationer, i mods\u00e6tning til at handle i den for at efterligne eller behage mennesker.<\/p>\n\n\n\n

Foto: CottonBro Studio<\/a><\/p>\n\n\n\n

H\u00f8r podcasten Fremtiden Forfra her om AI agenter<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"

Den generative AI-hype bobbel er ved at briste, s\u00e5 hvorfor ikke genindf\u00f8re en gammel id\u00e9…<\/p>\n","protected":false},"author":14,"featured_media":16508,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[108,2],"tags":[4260,4261,4262,4263,4264,4265],"class_list":["post-16510","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-nyt","category-nyheder","tag-agentic-ai-da","tag-ai-agents-da","tag-chatbots-da","tag-efficiency-gains-da","tag-generative-ai-da","tag-privacy-da-2"],"yoast_head":"\nAI hype i nye kl\u00e6der: AI agenter · Dataetisk Tænkehandletank<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/\" \/>\n<meta property=\"og:locale\" content=\"da_DK\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"AI hype i nye kl\u00e6der: AI agenter · Dataetisk Tænkehandletank\" \/>\n<meta property=\"og:description\" content=\"Den generative AI-hype bobbel er ved at briste, s\u00e5 hvorfor ikke genindf\u00f8re en gammel id\u00e9...\" \/>\n<meta property=\"og:url\" content=\"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/\" \/>\n<meta property=\"og:site_name\" content=\"Dataetisk Tænkehandletank\" \/>\n<meta property=\"article:published_time\" content=\"2025-08-26T08:37:24+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2025-08-26T12:50:43+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/dataethics.eu\/wp-content\/uploads\/2025\/08\/agenticAI.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"1000\" \/>\n\t<meta property=\"og:image:height\" content=\"666\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"Gry Hasselbalch & Pernille Tranberg\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@dataethicseu\" \/>\n<meta name=\"twitter:site\" content=\"@dataethicseu\" \/>\n<meta name=\"twitter:label1\" content=\"Skrevet af\" \/>\n\t<meta name=\"twitter:data1\" content=\"Gry Hasselbalch & Pernille Tranberg\" \/>\n\t<meta name=\"twitter:label2\" content=\"Estimeret l\u00e6setid\" \/>\n\t<meta name=\"twitter:data2\" content=\"7 minutter\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/the-ai-hype-in-a-new-guise-agentic-ai\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/the-ai-hype-in-a-new-guise-agentic-ai\\\/\"},\"author\":{\"name\":\"Gry Hasselbalch & Pernille Tranberg\",\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/#\\\/schema\\\/person\\\/e6d4e6f749abef1f2052f85d08b203a0\"},\"headline\":\"AI hype i nye kl\u00e6der: AI agenter\",\"datePublished\":\"2025-08-26T08:37:24+00:00\",\"dateModified\":\"2025-08-26T12:50:43+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/the-ai-hype-in-a-new-guise-agentic-ai\\\/\"},\"wordCount\":1354,\"publisher\":{\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/the-ai-hype-in-a-new-guise-agentic-ai\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/dataethics.eu\\\/wp-content\\\/uploads\\\/2025\\\/08\\\/agenticAI.jpg\",\"keywords\":[\"Agentic AI\",\"AI agents\",\"chatbots\",\"Efficiency gains\",\"generative AI\",\"privacy\"],\"articleSection\":[\"NYT\",\"Seneste nyheder\"],\"inLanguage\":\"da-DK\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/the-ai-hype-in-a-new-guise-agentic-ai\\\/\",\"url\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/the-ai-hype-in-a-new-guise-agentic-ai\\\/\",\"name\":\"AI hype i nye kl\u00e6der: AI agenter · Dataetisk Tænkehandletank\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/the-ai-hype-in-a-new-guise-agentic-ai\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/the-ai-hype-in-a-new-guise-agentic-ai\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/dataethics.eu\\\/wp-content\\\/uploads\\\/2025\\\/08\\\/agenticAI.jpg\",\"datePublished\":\"2025-08-26T08:37:24+00:00\",\"dateModified\":\"2025-08-26T12:50:43+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/the-ai-hype-in-a-new-guise-agentic-ai\\\/#breadcrumb\"},\"inLanguage\":\"da-DK\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/dataethics.eu\\\/da\\\/the-ai-hype-in-a-new-guise-agentic-ai\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"da-DK\",\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/the-ai-hype-in-a-new-guise-agentic-ai\\\/#primaryimage\",\"url\":\"https:\\\/\\\/dataethics.eu\\\/wp-content\\\/uploads\\\/2025\\\/08\\\/agenticAI.jpg\",\"contentUrl\":\"https:\\\/\\\/dataethics.eu\\\/wp-content\\\/uploads\\\/2025\\\/08\\\/agenticAI.jpg\",\"width\":1000,\"height\":666,\"caption\":\"Screenshot\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/the-ai-hype-in-a-new-guise-agentic-ai\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Hjem\",\"item\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"AI hype i nye kl\u00e6der: AI agenter\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/#website\",\"url\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/\",\"name\":\"Dataetisk T\u00e6nkehandletank\",\"description\":\"Alt om data-etik\",\"publisher\":{\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"da-DK\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/#organization\",\"name\":\"Data Ethics\",\"url\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"da-DK\",\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/dataethics.eu\\\/wp-content\\\/uploads\\\/2021\\\/04\\\/logo_yellow.png\",\"contentUrl\":\"https:\\\/\\\/dataethics.eu\\\/wp-content\\\/uploads\\\/2021\\\/04\\\/logo_yellow.png\",\"width\":124,\"height\":151,\"caption\":\"Data Ethics\"},\"image\":{\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/x.com\\\/dataethicseu\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/#\\\/schema\\\/person\\\/e6d4e6f749abef1f2052f85d08b203a0\",\"name\":\"Gry Hasselbalch & Pernille Tranberg\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"da-DK\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/648ce930182cec0befe2195f4f170ccb60a295dd3773282b3aaaadde0aaec125?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/648ce930182cec0befe2195f4f170ccb60a295dd3773282b3aaaadde0aaec125?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/648ce930182cec0befe2195f4f170ccb60a295dd3773282b3aaaadde0aaec125?s=96&d=mm&r=g\",\"caption\":\"Gry Hasselbalch & Pernille Tranberg\"},\"url\":\"https:\\\/\\\/dataethics.eu\\\/da\\\/author\\\/gry-og-pernille\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"AI hype i nye kl\u00e6der: AI agenter · Dataetisk Tænkehandletank","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/","og_locale":"da_DK","og_type":"article","og_title":"AI hype i nye kl\u00e6der: AI agenter · Dataetisk Tænkehandletank","og_description":"Den generative AI-hype bobbel er ved at briste, s\u00e5 hvorfor ikke genindf\u00f8re en gammel id\u00e9...","og_url":"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/","og_site_name":"Dataetisk Tænkehandletank","article_published_time":"2025-08-26T08:37:24+00:00","article_modified_time":"2025-08-26T12:50:43+00:00","og_image":[{"width":1000,"height":666,"url":"https:\/\/dataethics.eu\/wp-content\/uploads\/2025\/08\/agenticAI.jpg","type":"image\/jpeg"}],"author":"Gry Hasselbalch & Pernille Tranberg","twitter_card":"summary_large_image","twitter_creator":"@dataethicseu","twitter_site":"@dataethicseu","twitter_misc":{"Skrevet af":"Gry Hasselbalch & Pernille Tranberg","Estimeret l\u00e6setid":"7 minutter"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/#article","isPartOf":{"@id":"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/"},"author":{"name":"Gry Hasselbalch & Pernille Tranberg","@id":"https:\/\/dataethics.eu\/da\/#\/schema\/person\/e6d4e6f749abef1f2052f85d08b203a0"},"headline":"AI hype i nye kl\u00e6der: AI agenter","datePublished":"2025-08-26T08:37:24+00:00","dateModified":"2025-08-26T12:50:43+00:00","mainEntityOfPage":{"@id":"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/"},"wordCount":1354,"publisher":{"@id":"https:\/\/dataethics.eu\/da\/#organization"},"image":{"@id":"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/#primaryimage"},"thumbnailUrl":"https:\/\/dataethics.eu\/wp-content\/uploads\/2025\/08\/agenticAI.jpg","keywords":["Agentic AI","AI agents","chatbots","Efficiency gains","generative AI","privacy"],"articleSection":["NYT","Seneste nyheder"],"inLanguage":"da-DK"},{"@type":"WebPage","@id":"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/","url":"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/","name":"AI hype i nye kl\u00e6der: AI agenter · Dataetisk Tænkehandletank","isPartOf":{"@id":"https:\/\/dataethics.eu\/da\/#website"},"primaryImageOfPage":{"@id":"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/#primaryimage"},"image":{"@id":"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/#primaryimage"},"thumbnailUrl":"https:\/\/dataethics.eu\/wp-content\/uploads\/2025\/08\/agenticAI.jpg","datePublished":"2025-08-26T08:37:24+00:00","dateModified":"2025-08-26T12:50:43+00:00","breadcrumb":{"@id":"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/#breadcrumb"},"inLanguage":"da-DK","potentialAction":[{"@type":"ReadAction","target":["https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/"]}]},{"@type":"ImageObject","inLanguage":"da-DK","@id":"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/#primaryimage","url":"https:\/\/dataethics.eu\/wp-content\/uploads\/2025\/08\/agenticAI.jpg","contentUrl":"https:\/\/dataethics.eu\/wp-content\/uploads\/2025\/08\/agenticAI.jpg","width":1000,"height":666,"caption":"Screenshot"},{"@type":"BreadcrumbList","@id":"https:\/\/dataethics.eu\/da\/the-ai-hype-in-a-new-guise-agentic-ai\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Hjem","item":"https:\/\/dataethics.eu\/da\/"},{"@type":"ListItem","position":2,"name":"AI hype i nye kl\u00e6der: AI agenter"}]},{"@type":"WebSite","@id":"https:\/\/dataethics.eu\/da\/#website","url":"https:\/\/dataethics.eu\/da\/","name":"Dataetisk T\u00e6nkehandletank","description":"Alt om data-etik","publisher":{"@id":"https:\/\/dataethics.eu\/da\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/dataethics.eu\/da\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"da-DK"},{"@type":"Organization","@id":"https:\/\/dataethics.eu\/da\/#organization","name":"Data Ethics","url":"https:\/\/dataethics.eu\/da\/","logo":{"@type":"ImageObject","inLanguage":"da-DK","@id":"https:\/\/dataethics.eu\/da\/#\/schema\/logo\/image\/","url":"https:\/\/dataethics.eu\/wp-content\/uploads\/2021\/04\/logo_yellow.png","contentUrl":"https:\/\/dataethics.eu\/wp-content\/uploads\/2021\/04\/logo_yellow.png","width":124,"height":151,"caption":"Data Ethics"},"image":{"@id":"https:\/\/dataethics.eu\/da\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/x.com\/dataethicseu"]},{"@type":"Person","@id":"https:\/\/dataethics.eu\/da\/#\/schema\/person\/e6d4e6f749abef1f2052f85d08b203a0","name":"Gry Hasselbalch & Pernille Tranberg","image":{"@type":"ImageObject","inLanguage":"da-DK","@id":"https:\/\/secure.gravatar.com\/avatar\/648ce930182cec0befe2195f4f170ccb60a295dd3773282b3aaaadde0aaec125?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/648ce930182cec0befe2195f4f170ccb60a295dd3773282b3aaaadde0aaec125?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/648ce930182cec0befe2195f4f170ccb60a295dd3773282b3aaaadde0aaec125?s=96&d=mm&r=g","caption":"Gry Hasselbalch & Pernille Tranberg"},"url":"https:\/\/dataethics.eu\/da\/author\/gry-og-pernille\/"}]}},"_links":{"self":[{"href":"https:\/\/dataethics.eu\/da\/wp-json\/wp\/v2\/posts\/16510","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/dataethics.eu\/da\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/dataethics.eu\/da\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/dataethics.eu\/da\/wp-json\/wp\/v2\/users\/14"}],"replies":[{"embeddable":true,"href":"https:\/\/dataethics.eu\/da\/wp-json\/wp\/v2\/comments?post=16510"}],"version-history":[{"count":7,"href":"https:\/\/dataethics.eu\/da\/wp-json\/wp\/v2\/posts\/16510\/revisions"}],"predecessor-version":[{"id":16522,"href":"https:\/\/dataethics.eu\/da\/wp-json\/wp\/v2\/posts\/16510\/revisions\/16522"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/dataethics.eu\/da\/wp-json\/wp\/v2\/media\/16508"}],"wp:attachment":[{"href":"https:\/\/dataethics.eu\/da\/wp-json\/wp\/v2\/media?parent=16510"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/dataethics.eu\/da\/wp-json\/wp\/v2\/categories?post=16510"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/dataethics.eu\/da\/wp-json\/wp\/v2\/tags?post=16510"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}