spot_img
HomeApplications & Use CasesTV 2's KI-Kjetil: An AI Avatar's Impact on Election...

TV 2’s KI-Kjetil: An AI Avatar’s Impact on Election Journalism Recognized with European Award

TLDR: TV 2’s innovative AI-powered avatar, KI-Kjetil, has been honored with a silver award at the EBU Technology & Innovation Award 2025, recognizing it as one of Europe’s top media innovations. The digital replica of journalist Kjetil H. Dale was instrumental in transforming the broadcaster’s coverage of the American presidential election, making complex information more accessible to Norwegian audiences by answering over 70,000 questions with high accuracy.

TV 2, a prominent Norwegian broadcaster, has received significant international acclaim for its groundbreaking artificial intelligence (AI) avatar, ‘KI-Kjetil.’ The AI-driven news service was awarded silver in the EBU Technology & Innovation Award 2025, distinguishing it as one of Europe’s three leading media innovations. This recognition highlights TV 2’s successful integration of AI into journalism, particularly in its coverage of the American presidential election.

KI-Kjetil, a digital replica of TV 2 journalist Kjetil H. Dale, was developed with the primary goal of simplifying the complexities of the US election for the Norwegian public. The avatar served as an interactive news anchor, allowing users to pose questions about the election generally and TV 2’s exclusive average polls specifically. This personalized approach aimed to make political information more digestible and engaging for a broad audience, from those deeply invested in American politics to those seeking a basic understanding.

Utilizing advanced AI technologies, KI-Kjetil was meticulously crafted. Its synthetic voice was built from approximately 96 minutes of recordings of Dale’s actual voice using ElevenLabs, while its lifelike facial animation was created with Heygen, based on about three minutes of video footage of the journalist. Kjetil H. Dale himself was actively involved in the project, lending his likeness and voice and providing consent throughout the development process.

During its operational period from October 29 to November 6, KI-Kjetil was available for 106 hours and responded to an impressive 70,000 questions from the public. The avatar’s activity peaked during election day and night, demonstrating its utility during critical news cycles. Remarkably, TV 2 documented only seven errors in the avatar’s responses, underscoring its accuracy and reliability.

Chris Ronald Hermansen, Head of Editorial Development at TV 2 and project leader for editorial AI, expressed immense pride in the achievement. “It is really great to be noticed internationally for something we built from scratch. We ventured into something no one had a blueprint for, and the fact that we succeeded – together – makes me truly proud,” he stated. Hermansen also emphasized the importance of fostering an environment where innovation can thrive: “We actually got room to challenge – and that’s not something you should take for granted. This recognition shows what can happen when people are trusted to test boundaries and learn as they build.”

TV 2 maintained a strong commitment to journalistic ethics and transparency throughout the experimental project. While KI-Kjetil was designed to define its own responses, extensive pre-training ensured adherence to TV 2’s editorial guidelines. The avatar’s answers were strictly based on facts and source criticism, with journalists and researchers meticulously training it on objective information. Strict limitations were put in place to prevent the avatar from speculating or offering opinions. The project was transparently presented as an experiment and was removed after the election results were finalized. Despite some initial concerns about potential impacts on trust, TV 2’s regular trust measurements indicated no negative effect on public confidence in the broadcaster.

Also Read:

This pioneering initiative by TV 2 not only enhanced its USA election coverage and conveyed exclusive polling data in a novel way but also set a benchmark for integrating data journalism and technology responsibly. Hermansen affirmed TV 2’s ongoing commitment to AI, stating, “Our mission is to help make TV 2’s journalism more efficient and impactful using artificial intelligence.”

Karthik Mehta
Karthik Mehtahttps://blogs.edgentiq.com
Karthik Mehta is a data journalist known for his data-rich, insightful coverage of AI news and developments. Armed with a degree in Data Science from IIT Bombay and years of newsroom experience, Karthik merges storytelling with metrics to surface deeper narratives in AI-related events. His writing cuts through hype, revealing the real-world impact of Generative AI on industries, policy, and society. You can reach him out at: [email protected]

- Advertisement -

spot_img

Gen AI News and Updates

spot_img

- Advertisement -