Finnish Utopia Analytics’ AI technology receives and analyses millions of messages and images from clients including the Norwegian media group Schibsted, the Brazilian media UOL and Danish Momio.

Clients such as Momio formulate the guidelines that determine which messages and images Utopia Analytics’ AI technology should remove, explains CEO Dr. Mari-Sanna Paukkeri.

But for Utopia Analytics, the UN’s human rights also play a central role. The company doesn’t sell its AI technology to just anyone. In contracts with clients, for example, it is stipulated that both parties may cancel the partnership if human rights are not upheld.

“For example, there has been a discussion about differentiating on the basis of gender in clients’ guidelines. I’ve strongly advised against that, because it doesn’t promote gender equality. When our AI moderates messages and images, it should only look at the content itself and not at who it comes from,” says Mari-Sanna Paukkeri.

According to the CEO, the company has never ended a partnership due to human rights violations

Objective, precise and adaptable

It would require many employees if Utopia Analytics were to manually analyse and moderate all data the company receives.

That job is left to AI technology, and according to Mari-Sanna Paukkeri, it is much more precise and consistent than any human moderator.

“Humans are generally very subjective. Their decisions can vary from day to day and particularly from person to person. The AI moderator can do the same thing each day and even adapt along the way,” she says.

Powerful, but only a tool

Utopia Analytics’ AI-moderator can analyse text across all languages and all dialects. It can be used on any platform on which users generate a form of data. But even AI sometimes needs help.

“The technology is still only an aid for humans. If it comes across a challenge it cannot solve, it will always ask a human for help. In that way, algorithms are constantly developing, and AI technology gets smarter,” explains Mari-Sanna Paukkeri.