AI Firms Work to Give Chatbots a ‘European Perspective’

European AI firms want to make sure Google’s Bard and ChatGPT are speaking their language.

It’s why, as the Financial Times (FT) reported Wednesday (April 5), these artificial intelligence (AI) companies are working to improve the quality of responses from these chatbots in languages besides English.

Among them is Silo AI, a Finnish firm that is readying a project that would help build new large language models that would help the chatbots answer in languages such as Swedish, Danish and Norwegian.

As the FT notes, the rising popularity of AI software from companies like Google and OpenAI — backed by Microsoft— has critics worried about the use of a powerful technology kept in the hands of a small group of American participants.

“A European initiative needs to . . . capture knowledge from a European perspective and we can control what kind of data is being fed into it,” Peter Sarlin, chief executive of Silo AI, said in an interview with the FT.

Bard — for now — works only in English, the FT notes. And when ChatGPT supports several languages, the report says users who have done extensive tests on the chatbot have noticed that its accuracy does not hold up across all languages.

According to the FT, Silo AI is one of a number of groups trying to bolster the technology driving the chatbots with languages such as Arabic, Hebrew and German.

These projects are happening amid concurrent efforts by regulators to keep up with — and contain — AI’s entry into the global economy.

“As ChatGPT’s popularity has exploded, global concerns have been mounting about the AI industry’s lack of regulation, and lawmaker scrutiny around AI technology has increasingly shone a spotlight on an emerging regulation-innovation tug of war,” PYMNTS wrote recently.

For example, last week, Italy became the first Western country to outlaw OpenAI’s ChatGPT-4 after the nation’s Data Protection Authority announced an investigation of the chatbot’s alleged breach of General Data Protection Regulation privacy rules and age-verification practices.

The technology has already been blocked by China, Russia, North Korea and Iran.

The Italian regulator claimed that there was “an absence of any legal basis” justifying the massive data collection and storage of personal information used to “train” the GPT-4 chatbot, and issued a temporary ban on its use in the country.

Italy’s move has prompted other privacy regulators in Europe to take a closer look at AI tools, with regulators in Ireland, Germany and France reportedly considering their own investigations.

For all PYMNTS EMEA coverage, subscribe to the daily EMEA Newsletter.