Department of Information Science and Media Studies

- Careful feeding the AI beast

- As long as we don't know more about how Big Tech use the information we share with them, we should be careful about feeding the AI technology with potentially sensitive material, writes experts from the department in an op-ed published in Medier24.

@julientromeur via Unsplash

Main content

Thanks to its exceptional ability to generate human-like text, ChatGPT has become a popular tool among many users, including journalists. The AI-powered tool, however, carries inherent risks that should not be overlooked.

In an op-ed featured in Medier 24, our natural language processing researcher Samia Touileb and Per Christian Magnus, the leader of the Centre for Investigative Journalism, shed light on the hazards associated with inputting potentially sensitive data into such large language models chatbots. The authors urge journalists and editors to exercise caution and critical thinking in their use of AI-technology.

Touileb is a researcher on Natural Language Processing and AI at MediaFutures, while head of SUJO Magnus has several years of  experience as reporter, journalist and editor at TV2 and NRK.

- This op-ed is a good example of the width of Infomedia's compentence, says head of department Marija Slavkovik about the recently publised text.

Read the op-ed (in Norwegian) in question here.