{"id":176512,"date":"2025-05-02T04:07:56","date_gmt":"2025-05-02T02:07:56","guid":{"rendered":"https:\/\/uniavisen.dk\/kunstig-intelligens-kan-aendre-spillereglerne-i-asylretten\/"},"modified":"2025-05-02T07:43:08","modified_gmt":"2025-05-02T05:43:08","slug":"new-ai-model-to-transform-how-asylum-cases-are-judged","status":"publish","type":"post","link":"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/","title":{"rendered":"New AI model to transform how asylum cases are judged"},"content":{"rendered":"

Lawyers and caseworkers in Denmark will soon be able to use AI when making decisions in asylum cases.<\/p>\n

This is the purpose of a five-year project titled Explainable Artificial Intelligence and Credibility in Asylum Decision-making<\/em>, led by two professors.<\/p>\n

One is Professor of Law Thomas Gammeltoft-Hansen at the Nordic Asylum Law and Data Lab<\/a> and head of the Centre of Excellence for Global Mobility Law<\/a> at the University of Copenhagen (UCPH). The other is Thomas Moeslund, one of the world\u2019s leading researchers in ‘Explainable AI’<\/span>, based at Aalborg University. Explainable AI is a set of tools and frameworks that make artificial intelligence systems transparent and comprehensible.<\/p>\n

Together, they have secured DKK 12 million from the Villum Foundation to develop an AI that, according to Thomas Gammeltoft-Hansen, could open up entirely new options for support of asylum decision-making.<\/p>\n

\u00bbOur AI is designed to provide transparency on how caseworkers or judges arrive at the most legally sound decisions in asylum cases. We also hope that newly hired lawyers and attorneys can use it as a training tool,\u00ab says Thomas Gammeltoft-Hansen.<\/p>\n

The artificial intelligence is not intended to replace<\/em> human decision-making, but to support lawyers and caseworkers \u2014 and to minimise bias in assessments.<\/p>\n

Bias is a major challenge in AI<\/h4>\n

From his office on South Campus, Thomas Gammeltoft-Hansen explains that a core issue in asylum law is the verification of the claims of asylum seekers: the outcome often hinges entirely on the applicant\u2019s credibility.<\/p>\n

\u00bbThe practice of processing asylum cases varies greatly from country to country, and there are no shared standards for assessing credibility. This makes it easy for individual, systemic, and institutional bias<\/span> to creep in,\u00ab says Thomas Gammeltoft-Hansen.<\/p>\n

AI is trained on past decisions, and this introduces the risk of reproducing and reinforcing existing biases and distortions. For this reason, Thomas Gammeltoft-Hansen does not believe that AI can be used to make independent asylum decisions in an ethically responsible manner \u2014 or certainly not at present.<\/p>\n

A study from Canada<\/a> shows that the rate of granted residence permits ranged from 13.8 to 95.1 per cent, depending on the judge assigned to the case.<\/p>\n

\n

PROFILE<\/p>\n

Thomas Gammeltoft-Hansen <\/strong>has a master\u2019s degree in refugee studies from the University of Oxford (2003), a degree in political science from the University of Copenhagen (2005), and a PhD from Aarhus University (2009).<\/p>\n

After completing his PhD, he was hired by the Danish Institute for International Studies (DIIS) in 2009.<\/p>\n

From 2013, he served as research director at the Danish Institute for Human Rights and, during the same period, was appointed member of the Danish Refugee Appeals Board.<\/p>\n

From 2016, he was research director at the Raoul Wallenberg Institute before joining UCPH in 2018<\/strong>.<\/p>\n<\/div>\n

According to Thomas Gammeltoft-Hansen, the problem is less pronounced in Denmark. There is no gender bias here, as seen in other countries, but religion can play a larger role in case outcomes. These are precisely the kinds of underlying \u2014 and often invisible \u2014 patterns the project aims to investigate and expose.<\/p>\n

\u00bbAsylum law almost always involves a subjective assessment of how much the caseworker believes and weighs the applicant\u2019s explanation. One way we can work with this issue is to acknowledge that the subjective element is ever-present. When training our AI, we have to recognise that we are working with a dataset in which bias is constantly embedded,\u00ab says Thomas Gammeltoft-Hansen.<\/p>\n

AI for asylum must be subject to review<\/h4>\n

According to Thomas Gammeltoft-Hansen, he and Thomas Moeslund are using different technologies to address the challenges \u2014 and have already tested a pilot model.<\/p>\n

\u00bbBased on the pilot results, I feel confident saying our AI model is world-leading in terms of its ability to accurately predict asylum decisions,\u00ab he says.<\/p>\n

To achieve this, they had to use the latest developments in artificial intelligence:<\/p>\n

\u00bbThomas and I have worked to make AI models more transparent and controllable, but our field is still relatively new in the AI world. Most recent breakthroughs in AI have prioritised predictive power or output quality at the expense of understanding how algorithms arrive at their conclusions,\u00ab says Thomas Gammeltoft-Hansen.<\/p>\n

\u00bbThis phenomenon is often referred to as the black box problem in AI. But with the Explainable AI approach we use, we are going in a different direction: here, the algorithms help open the black box and give users insight into how large language models reach their conclusions,\u00ab he says.<\/p>\n

In legal contexts, this approach is absolutely necessary, according to Thomas Gammeltoft-Hansen.<\/p>\n

\u00bbWhen using algorithms to process sensitive personal data, or when AI supports administrative or legal decisions, there must be an individual and well-reasoned justification. So, opening the black box is crucial,\u00ab he says.<\/p>\n

Secured access to unique datasets<\/h4>\n

Thomas Gammeltoft-Hansen has always been interested in how legal practices evolve over time. That\u2019s also what drew his attention to the growing use of AI systems \u2014 especially in the United States.<\/p>\n

But to conduct research in this area, he needed access to a large dataset of asylum decisions \u2014 something that is notoriously hard to obtain in an international context.<\/p>\n

\u00bbTogether with Nordic colleagues, we gained access to extensive datasets from the Danish Refugee Appeals Board, the Norwegian Directorate of Immigration, and the Swedish Migration Agency. Today, we have access to over 800,000 decisions, and our data is truly unique,\u00ab says Thomas Gammeltoft-Hansen.<\/p>\n

What makes it unique is that these are so-called deep datasets. They are not merely short summaries of decisions \u2014 they include information on what asylum seekers were asked during processing, how they responded, and what arguments the attorneys presented, according to\u00a0Thomas Gammeltoft-Hansen.<\/p>\n

\u00bbIt\u2019s only because of our strong collaboration in the Nordic region that this has been possible. It has taken a long time to build trust and demonstrate that we can handle data securely and responsibly. This kind of access would not have been possible in many other parts of the world\u00ab.
\n<\/p>\n","protected":false},"excerpt":{"rendered":"

In a few years time, and with an AI model under development at the Faculty of Law, we can predict how religion, nationality or education can influence the outcome of an asylum case. The aim is to prevent bias in case processing.<\/p>\n","protected":false},"author":8,"featured_media":176552,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"om_disable_all_campaigns":false,"_monsterinsights_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0,"_uf_show_specific_survey":0,"_uf_disable_surveys":false,"footnotes":""},"categories":[46],"tags":[],"class_list":["post-176512","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-science","expression-news_article"],"acf":[],"aioseo_notices":[],"yoast_head":"\nNew AI model to transform how asylum cases are judged \u2014 University Post<\/title>\n<meta name=\"description\" content=\"Lawyers and caseworkers will in the future be able to use AI to help them make the best possible decisions in asylum cases.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"New AI model to transform how asylum cases are judged \u2014 University Post\" \/>\n<meta property=\"og:description\" content=\"Lawyers and caseworkers will in the future be able to use AI to help them make the best possible decisions in asylum cases.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/\" \/>\n<meta property=\"og:site_name\" content=\"University Post\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/uniavis\" \/>\n<meta property=\"article:published_time\" content=\"2025-05-02T02:07:56+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2025-05-02T05:43:08+00:00\" \/>\n<meta property=\"og:image\" content=\"http:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansen04.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"1600\" \/>\n\t<meta property=\"og:image:height\" content=\"1068\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"Anders Fjeldberg\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:image\" content=\"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansen04.jpg\" \/>\n<meta name=\"twitter:creator\" content=\"@Uniavisen\" \/>\n<meta name=\"twitter:site\" content=\"@Uniavisen\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Anders Fjeldberg\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"5 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/uniavisen.dk\\\/en\\\/new-ai-model-to-transform-how-asylum-cases-are-judged\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/uniavisen.dk\\\/en\\\/new-ai-model-to-transform-how-asylum-cases-are-judged\\\/\"},\"author\":{\"name\":\"Anders Fjeldberg\",\"@id\":\"https:\\\/\\\/uniavisen.dk\\\/#\\\/schema\\\/person\\\/e92ba15746694abe6835e7ad00240ca7\"},\"headline\":\"New AI model to transform how asylum cases are judged\",\"datePublished\":\"2025-05-02T02:07:56+00:00\",\"dateModified\":\"2025-05-02T05:43:08+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/uniavisen.dk\\\/en\\\/new-ai-model-to-transform-how-asylum-cases-are-judged\\\/\"},\"wordCount\":978,\"image\":{\"@id\":\"https:\\\/\\\/uniavisen.dk\\\/en\\\/new-ai-model-to-transform-how-asylum-cases-are-judged\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/uniavisen.dk\\\/wp-content\\\/uploads\\\/2025\\\/04\\\/thomasgammeltofthansentilfacebook-1.jpg\",\"articleSection\":[\"Science\"],\"inLanguage\":\"en-US\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/uniavisen.dk\\\/en\\\/new-ai-model-to-transform-how-asylum-cases-are-judged\\\/\",\"url\":\"https:\\\/\\\/uniavisen.dk\\\/en\\\/new-ai-model-to-transform-how-asylum-cases-are-judged\\\/\",\"name\":\"New AI model to transform how asylum cases are judged \u2014 University Post\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/uniavisen.dk\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/uniavisen.dk\\\/en\\\/new-ai-model-to-transform-how-asylum-cases-are-judged\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/uniavisen.dk\\\/en\\\/new-ai-model-to-transform-how-asylum-cases-are-judged\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/uniavisen.dk\\\/wp-content\\\/uploads\\\/2025\\\/04\\\/thomasgammeltofthansentilfacebook-1.jpg\",\"datePublished\":\"2025-05-02T02:07:56+00:00\",\"dateModified\":\"2025-05-02T05:43:08+00:00\",\"author\":{\"@id\":\"https:\\\/\\\/uniavisen.dk\\\/#\\\/schema\\\/person\\\/e92ba15746694abe6835e7ad00240ca7\"},\"description\":\"Lawyers and caseworkers will in the future be able to use AI to help them make the best possible decisions in asylum cases.\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/uniavisen.dk\\\/en\\\/new-ai-model-to-transform-how-asylum-cases-are-judged\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/uniavisen.dk\\\/en\\\/new-ai-model-to-transform-how-asylum-cases-are-judged\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/uniavisen.dk\\\/en\\\/new-ai-model-to-transform-how-asylum-cases-are-judged\\\/#primaryimage\",\"url\":\"https:\\\/\\\/uniavisen.dk\\\/wp-content\\\/uploads\\\/2025\\\/04\\\/thomasgammeltofthansentilfacebook-1.jpg\",\"contentUrl\":\"https:\\\/\\\/uniavisen.dk\\\/wp-content\\\/uploads\\\/2025\\\/04\\\/thomasgammeltofthansentilfacebook-1.jpg\",\"width\":1200,\"height\":674},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/uniavisen.dk\\\/en\\\/new-ai-model-to-transform-how-asylum-cases-are-judged\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/uniavisen.dk\\\/en\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"New AI model to transform how asylum cases are judged\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/uniavisen.dk\\\/#website\",\"url\":\"https:\\\/\\\/uniavisen.dk\\\/\",\"name\":\"University Post\",\"description\":\"Independent of management\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/uniavisen.dk\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/uniavisen.dk\\\/#\\\/schema\\\/person\\\/e92ba15746694abe6835e7ad00240ca7\",\"name\":\"Anders Fjeldberg\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/dbd29cb1d1ebf4635c343a9e2606601d58f406394ebf2f41b2d976b72071e42f?s=96&d=identicon&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/dbd29cb1d1ebf4635c343a9e2606601d58f406394ebf2f41b2d976b72071e42f?s=96&d=identicon&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/dbd29cb1d1ebf4635c343a9e2606601d58f406394ebf2f41b2d976b72071e42f?s=96&d=identicon&r=g\",\"caption\":\"Anders Fjeldberg\"},\"url\":\"https:\\\/\\\/uniavisen.dk\\\/en\\\/author\\\/anders\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"New AI model to transform how asylum cases are judged \u2014 University Post","description":"Lawyers and caseworkers will in the future be able to use AI to help them make the best possible decisions in asylum cases.","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:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/","og_locale":"en_US","og_type":"article","og_title":"New AI model to transform how asylum cases are judged \u2014 University Post","og_description":"Lawyers and caseworkers will in the future be able to use AI to help them make the best possible decisions in asylum cases.","og_url":"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/","og_site_name":"University Post","article_publisher":"https:\/\/www.facebook.com\/uniavis","article_published_time":"2025-05-02T02:07:56+00:00","article_modified_time":"2025-05-02T05:43:08+00:00","og_image":[{"width":1600,"height":1068,"url":"http:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansen04.jpg","type":"image\/jpeg"}],"author":"Anders Fjeldberg","twitter_card":"summary_large_image","twitter_image":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansen04.jpg","twitter_creator":"@Uniavisen","twitter_site":"@Uniavisen","twitter_misc":{"Written by":"Anders Fjeldberg","Est. reading time":"5 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/#article","isPartOf":{"@id":"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/"},"author":{"name":"Anders Fjeldberg","@id":"https:\/\/uniavisen.dk\/#\/schema\/person\/e92ba15746694abe6835e7ad00240ca7"},"headline":"New AI model to transform how asylum cases are judged","datePublished":"2025-05-02T02:07:56+00:00","dateModified":"2025-05-02T05:43:08+00:00","mainEntityOfPage":{"@id":"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/"},"wordCount":978,"image":{"@id":"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/#primaryimage"},"thumbnailUrl":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansentilfacebook-1.jpg","articleSection":["Science"],"inLanguage":"en-US"},{"@type":"WebPage","@id":"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/","url":"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/","name":"New AI model to transform how asylum cases are judged \u2014 University Post","isPartOf":{"@id":"https:\/\/uniavisen.dk\/#website"},"primaryImageOfPage":{"@id":"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/#primaryimage"},"image":{"@id":"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/#primaryimage"},"thumbnailUrl":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansentilfacebook-1.jpg","datePublished":"2025-05-02T02:07:56+00:00","dateModified":"2025-05-02T05:43:08+00:00","author":{"@id":"https:\/\/uniavisen.dk\/#\/schema\/person\/e92ba15746694abe6835e7ad00240ca7"},"description":"Lawyers and caseworkers will in the future be able to use AI to help them make the best possible decisions in asylum cases.","breadcrumb":{"@id":"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/#primaryimage","url":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansentilfacebook-1.jpg","contentUrl":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansentilfacebook-1.jpg","width":1200,"height":674},{"@type":"BreadcrumbList","@id":"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/uniavisen.dk\/en\/"},{"@type":"ListItem","position":2,"name":"New AI model to transform how asylum cases are judged"}]},{"@type":"WebSite","@id":"https:\/\/uniavisen.dk\/#website","url":"https:\/\/uniavisen.dk\/","name":"University Post","description":"Independent of management","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/uniavisen.dk\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Person","@id":"https:\/\/uniavisen.dk\/#\/schema\/person\/e92ba15746694abe6835e7ad00240ca7","name":"Anders Fjeldberg","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/dbd29cb1d1ebf4635c343a9e2606601d58f406394ebf2f41b2d976b72071e42f?s=96&d=identicon&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/dbd29cb1d1ebf4635c343a9e2606601d58f406394ebf2f41b2d976b72071e42f?s=96&d=identicon&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/dbd29cb1d1ebf4635c343a9e2606601d58f406394ebf2f41b2d976b72071e42f?s=96&d=identicon&r=g","caption":"Anders Fjeldberg"},"url":"https:\/\/uniavisen.dk\/en\/author\/anders\/"}]}},"advancedCustomFields":{"expression":{"term_id":15,"name":"News Article","slug":"news_article","term_group":0,"term_taxonomy_id":15,"taxonomy":"expression","description":"","parent":0,"count":11489,"filter":"raw"},"enable_comments":true,"align_content":"alignleft","feature_color":"","article_updated":"","layout_group":[{"acf_fc_layout":"Headline","use_post_title":true,"headline":"","style":"default","highlighted_words":"","text_size":"small"},{"acf_fc_layout":"Image","image":{"ID":176583,"id":176583,"title":"Thomas Gammeltoft-Hansen04","filename":"thomasgammeltofthansen04.jpg","filesize":150200,"url":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansen04.jpg","link":"https:\/\/uniavisen.dk\/en\/new-ai-model-to-transform-how-asylum-cases-are-judged\/thomas-gammeltoft-hansen04-2\/","alt":"","author":"8","description":"","caption":"Juraprofessor Thomas Gammeltoft-Hansen er den ene af to forskere, der er ved at udvikle en Explainable AI-model, der skal fjerne fejl og usikkerheder i asylafg\u00f8relser.","name":"thomas-gammeltoft-hansen04-2","status":"inherit","uploaded_to":176512,"date":"2025-05-01 08:10:39","modified":"2025-10-16 08:32:45","menu_order":0,"mime_type":"image\/jpeg","type":"image","subtype":"jpeg","icon":"https:\/\/uniavisen.dk\/wp-includes\/images\/media\/default.png","width":1600,"height":1068,"sizes":{"thumbnail":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansen04-150x150.jpg","thumbnail-width":150,"thumbnail-height":150,"medium":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansen04-480x320.jpg","medium-width":480,"medium-height":320,"medium_large":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansen04-768x513.jpg","medium_large-width":768,"medium_large-height":513,"large":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansen04-1280x854.jpg","large-width":1280,"large-height":854,"1536x1536":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansen04-1536x1025.jpg","1536x1536-width":1536,"1536x1536-height":1025,"2048x2048":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansen04.jpg","2048x2048-width":1600,"2048x2048-height":1068,"featured-soft":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansen04-290x194.jpg","featured-soft-width":290,"featured-soft-height":194,"featured-hard":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansen04-290x180.jpg","featured-hard-width":290,"featured-hard-height":180,"narrow":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansen04-700x467.jpg","narrow-width":700,"narrow-height":467,"extended":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansen04-990x661.jpg","extended-width":990,"extended-height":661}},"style":"extended","text_placement":"metadata-below","image_link_url":"","image_link_title":"","caption_prefix":"","enable_alternative_caption":true,"alternative_caption":"Professor of Law Thomas Gammeltoft-Hansen"},{"acf_fc_layout":"Standfirst","subject":"Transparency","text":"Kunstig intelligens kan forudsige, hvordan religion, nationalitet eller uddannelse kan p\u00e5virke udfaldet af en asylans\u00f8gning, s\u00e5 nu er en AI-model, der skal hj\u00e6lpe sagsbehandlere med at undg\u00e5 bias, under udvikling p\u00e5 Det Juridiske fakultet.","use_post_excerpt":true},{"acf_fc_layout":"Byline","is_author":true,"contributors":false},{"acf_fc_layout":"Content","content":"<p>Lawyers and caseworkers in Denmark will soon be able to use AI when making decisions in asylum cases.<\/p>\n<p>This is the purpose of a five-year project titled <em>Explainable Artificial Intelligence and Credibility in Asylum Decision-making<\/em>, led by two professors.<\/p>\n<p>One is Professor of Law Thomas Gammeltoft-Hansen at the <a href=\"https:\/\/asylumdata.ku.dk\/\" target=\"_blank\" rel=\"noopener\">Nordic Asylum Law and Data Lab<\/a> and head of the <a href=\"https:\/\/dg.dk\/en\/centers\/center-for-global-mobility-law-mobile\/\" target=\"_blank\" rel=\"noopener\">Centre of Excellence for Global Mobility Law<\/a> at the University of Copenhagen (UCPH). The other is Thomas Moeslund, one of the world\u2019s leading researchers in ‘<span title=\"A set of tools and frameworks to make artificial intelligence systems transparent and comprehensible\">Explainable AI’<\/span>, based at Aalborg University. Explainable AI is a set of tools and frameworks that make artificial intelligence systems transparent and comprehensible.<\/p>\n<p>Together, they have secured DKK 12 million from the Villum Foundation to develop an AI that, according to Thomas Gammeltoft-Hansen, could open up entirely new options for support of asylum decision-making.<\/p>\n<p>\u00bbOur AI is designed to provide transparency on how caseworkers or judges arrive at the most legally sound decisions in asylum cases. We also hope that newly hired lawyers and attorneys can use it as a training tool,\u00ab says Thomas Gammeltoft-Hansen.<\/p>\n<p>The artificial intelligence is not intended to <em>replace<\/em> human decision-making, but to support lawyers and caseworkers \u2014 and to minimise bias in assessments.<\/p>\n<h4>Bias is a major challenge in AI<\/h4>\n<p>From his office on South Campus, Thomas Gammeltoft-Hansen explains that a core issue in asylum law is the verification of the claims of asylum seekers: the outcome often hinges entirely on the applicant\u2019s credibility.<\/p>\n<p>\u00bbThe practice of processing asylum cases varies greatly from country to country, and there are no shared standards for assessing credibility. This makes it easy for individual, systemic, and institutional <span title=\"Such as distortion, stereotypes, or prejudices that can influence decisions\">bias<\/span> to creep in,\u00ab says Thomas Gammeltoft-Hansen.<\/p>\n<p>AI is trained on past decisions, and this introduces the risk of reproducing and reinforcing existing biases and distortions. For this reason, Thomas Gammeltoft-Hansen does not believe that AI can be used to make independent asylum decisions in an ethically responsible manner \u2014 or certainly not at present.<\/p>\n<p>A <a href=\"https:\/\/www.administrativelawmatters.com\/blog\/2018\/09\/21\/sean-rehaag-on-revisiting-the-luck-of-the-draw\/\" target=\"_blank\" rel=\"noopener\">study from Canada<\/a> shows that the rate of granted residence permits ranged from 13.8 to 95.1 per cent, depending on the judge assigned to the case.<\/p>\n<div class=\"factbox\">\n<p class=\"factbox-header feature-color\">PROFILE<\/p>\n<p><strong>Thomas Gammeltoft-Hansen <\/strong>has a master\u2019s degree in refugee studies from the University of Oxford (2003), a degree in political science from the University of Copenhagen (2005), and a PhD from Aarhus University (2009).<\/p>\n<p>After completing his PhD, he was hired by the Danish Institute for International Studies (DIIS) in 2009.<\/p>\n<p>From 2013, he served as research director at the Danish Institute for Human Rights and, during the same period, was appointed member of the Danish Refugee Appeals Board.<\/p>\n<p>From 2016, he was research director at the Raoul Wallenberg Institute before joining <strong>UCPH in 2018<\/strong>.<\/p>\n<\/div>\n<p>According to Thomas Gammeltoft-Hansen, the problem is less pronounced in Denmark. There is no gender bias here, as seen in other countries, but religion can play a larger role in case outcomes. These are precisely the kinds of underlying \u2014 and often invisible \u2014 patterns the project aims to investigate and expose.<\/p>\n<p>\u00bbAsylum law almost always involves a subjective assessment of how much the caseworker believes and weighs the applicant\u2019s explanation. One way we can work with this issue is to acknowledge that the subjective element is ever-present. When training our AI, we have to recognise that we are working with a dataset in which bias is constantly embedded,\u00ab says Thomas Gammeltoft-Hansen.<\/p>\n<h4>AI for asylum must be subject to review<\/h4>\n<p>According to Thomas Gammeltoft-Hansen, he and Thomas Moeslund are using different technologies to address the challenges \u2014 and have already tested a pilot model.<\/p>\n<p>\u00bbBased on the pilot results, I feel confident saying our AI model is world-leading in terms of its ability to accurately predict asylum decisions,\u00ab he says.<\/p>\n<p>To achieve this, they had to use the latest developments in artificial intelligence:<\/p>\n<p>\u00bbThomas and I have worked to make AI models more transparent and controllable, but our field is still relatively new in the AI world. Most recent breakthroughs in AI have prioritised predictive power or output quality at the expense of understanding how algorithms arrive at their conclusions,\u00ab says Thomas Gammeltoft-Hansen.<\/p>\n<p>\u00bbThis phenomenon is often referred to as the black box problem in AI. But with the Explainable AI approach we use, we are going in a different direction: here, the algorithms help open the black box and give users insight into how large language models reach their conclusions,\u00ab he says.<\/p>\n<p>In legal contexts, this approach is absolutely necessary, according to Thomas Gammeltoft-Hansen.<\/p>\n<p>\u00bbWhen using algorithms to process sensitive personal data, or when AI supports administrative or legal decisions, there must be an individual and well-reasoned justification. So, opening the black box is crucial,\u00ab he says.<\/p>\n<h4>Secured access to unique datasets<\/h4>\n<p>Thomas Gammeltoft-Hansen has always been interested in how legal practices evolve over time. That\u2019s also what drew his attention to the growing use of AI systems \u2014 especially in the United States.<\/p>\n<p>But to conduct research in this area, he needed access to a large dataset of asylum decisions \u2014 something that is notoriously hard to obtain in an international context.<\/p>\n<p>\u00bbTogether with Nordic colleagues, we gained access to extensive datasets from the Danish Refugee Appeals Board, the Norwegian Directorate of Immigration, and the Swedish Migration Agency. Today, we have access to over 800,000 decisions, and our data is truly unique,\u00ab says Thomas Gammeltoft-Hansen.<\/p>\n<p>What makes it unique is that these are so-called deep datasets. They are not merely short summaries of decisions \u2014 they include information on what asylum seekers were asked during processing, how they responded, and what arguments the attorneys presented, according to\u00a0Thomas Gammeltoft-Hansen.<\/p>\n<p>\u00bbIt\u2019s only because of our strong collaboration in the Nordic region that this has been possible. It has taken a long time to build trust and demonstrate that we can handle data securely and responsibly. This kind of access would not have been possible in many other parts of the world\u00ab.<\/p>\n"},{"acf_fc_layout":"ArticleEnd"},{"acf_fc_layout":"Newsletter","lang_select":"en","identifier":"Newsletter","headline":"Get an email with upcoming events and top University of Copenhagen stories","button_text":"Sign up here","class":""},{"acf_fc_layout":"OtherStories","headline":"","hand_picked_posts":false,"references":false,"category":false,"theme":false,"number_of_posts":"4","style":"default"}]},"taxonomyData":{"category":[{"term_id":46,"name":"Science","slug":"science","term_group":0,"term_taxonomy_id":46,"taxonomy":"category","description":"","parent":0,"count":831,"filter":"raw"}],"post_tag":[],"post_format":[],"expression":[{"term_id":15,"name":"News Article","slug":"news_article","term_group":0,"term_taxonomy_id":15,"taxonomy":"expression","description":"","parent":0,"count":11489,"filter":"raw"}],"translation_priority":[{"term_id":5468,"name":"Optional","slug":"optional-en","term_group":0,"term_taxonomy_id":5468,"taxonomy":"translation_priority","description":"","parent":0,"count":672,"filter":"raw"}]},"featured_media_url":"https:\/\/uniavisen.dk\/wp-content\/uploads\/2025\/04\/thomasgammeltofthansentilfacebook-1.jpg","_links":{"self":[{"href":"https:\/\/uniavisen.dk\/en\/wp-json\/wp\/v2\/posts\/176512","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/uniavisen.dk\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/uniavisen.dk\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/uniavisen.dk\/en\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/uniavisen.dk\/en\/wp-json\/wp\/v2\/comments?post=176512"}],"version-history":[{"count":11,"href":"https:\/\/uniavisen.dk\/en\/wp-json\/wp\/v2\/posts\/176512\/revisions"}],"predecessor-version":[{"id":176611,"href":"https:\/\/uniavisen.dk\/en\/wp-json\/wp\/v2\/posts\/176512\/revisions\/176611"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/uniavisen.dk\/en\/wp-json\/wp\/v2\/media\/176552"}],"wp:attachment":[{"href":"https:\/\/uniavisen.dk\/en\/wp-json\/wp\/v2\/media?parent=176512"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/uniavisen.dk\/en\/wp-json\/wp\/v2\/categories?post=176512"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/uniavisen.dk\/en\/wp-json\/wp\/v2\/tags?post=176512"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}