18 December 2019

Finnish AI house Utopia Analytics has published a framework for ethically sustainable artificial intelligence. Utopia’s Ethical AI Manifesto is a set of guidelines that emphasize common principles such as transparency, human-first, comprehensibility, non-discrimination and human dignity.

Utopia Analytics is a world leading text analytics company. Its AI tools are able to moderate, quality-check and curate any type of user generated content in any language or dialect in the world. Utopia’s tools are used in automated handling of i.e. news comments, chat messages, online marketplace ads and insurance claims.

“AI is a tool, and humans are the master. People should always define what is right and what is wrong. Ethics comes from humans. But we at Utopia understand that AI can be a powerful tool. From the very beginning we have chosen the path of Ethical AI. As forerunners, we put these principles into action together with our customers,” says Utopia’s CEO Dr. Mari-Sanna Paukkeri.

Utopia’s Ethical AI is based on the United Nations Universal Declaration of Human Rights (UDHR). Utopia’s AI service contracts stipulate that if a party breaches UDHR, the contract will be terminated. This anchoring has been in the Utopia AI Moderator contracts since the beginning of Utopia Analytics. According to Paukkeri, the clause is important because it guides both parties to watch each other.

“Our AI works in all languages and dialects in the world, and our service is so scalable that it could moderate everywhere in the internet in real-time. Therefore, it needs a solid base. For that, we chose UN Universal Declaration of Human Rights. We never define what is OK for publishing and what is not. That is our customers’ privilege on their own platform. Every online service we moderate has their own policy. That is democracy” Paukkeri says.

In Ethical AI Manifesto, Utopia pledges not to sell its services to anyone who might abuse them, whilst at the same time taking full responsibility of its own algorithms and AI in general. Paukkeri stresses that Utopia focus on building unbiased AI models, models that moderate every message, image, claim or ad, solely by the content and context, no matter who created it.

“AI and ethics have become a major topic in the public discourse since so many tasks have been automated, and there is a mistaken belief that no one would be responsible for AI decisions. This is not true: the people who build AI models and who provide the training data for AI should always be responsible for the decisions the model makes” Paukkeri adds.

About Utopia Analytics

Utopia Analytics is a Finnish AI company, which specialises in analysing text in any language or dialect. The Helsinki-based company was founded in 2014.

Utopia AI Moderator and Utopia AI Claim Handler are the flagship products of the company. Both products can automatically moderate user-generated content in any language using Artificial Intelligence and machine learning. They are used successfully for example on the children’s Facebook-equivalent Momio, on the largest discussion forum in Finland Suomi24.fi, and on popular online marketplaces such as Swiss Tutti.ch as well as Finnish Tori.fi. Global frontline news channels such as Finnish Iltalehti’s and Brazilian UOL’s news comments are also moderated by Utopia.