European Companies Enter the AI Race

Artificial Intelligence (AI) has increasingly become a part of daily life for several years. ChatGPT, a conversational robot, marked a significant shift in how people perceive AI.

Many of the widely known AI chats are products of tech companies in the United States. Nevertheless, European firms have been concentrating more on business applications. They are optimistic that they will not be left behind in the fast-evolving field of AI.

European Bots

Numerous European startups are currently focused on creating their bots utilizing existing AI models. Others are developing their models. These startups are leveraging open-source code, similar to their US counterparts, to construct the knowledge base for their bots.

The German-based startup Aleph Alpha is working on a multimedia chatbot. Meanwhile, France’s Hugging Face partnered with France’s national scientific research center, CNRS, in coordinating researchers to build Bloom. Bloom is a scientific language model. Hugging Face has also recently partnered with Amazon.

Companies like Stable Diffusion are focusing on more specialized bots. It released its text-to-image model last year.

Despite their progress, these startups have raised only tens of millions of dollars from investors, leaving them far behind the billions spent by their US counterparts.

Competitive Tech

Despite the vast funding gap between European and US firms, European companies remain optimistic about their competitiveness and ability to capture a portion of the growing market.

Aleph Alpha believes that the firm’s latest bot can compete with OpenAI’s latest GPT-3 model. Aleph Alpha’s model uniquely utilizes text and images as input prompts. This creates opportunities for a wide range of applications.

Focusing on high-value work for corporations, Aleph Alpha is competing against Microsoft, and Andrulis acknowledged the need to move quickly.

According to Laurent Daudet, one of the co-founders of LightOn, a company based in Paris, their models functions like GPT-3.

Multilingual

A significant number of European AI models are designed to be multilingual. Bloom’s model can operate in seven languages, including Chinese, English, French, and Spanish. Also, Aleph Alpha’s model can operate in five languages. The multilingual feature helps these firms as Europeans prefer tools in their native languages. However, the quality of these models depends on the volume of text fed into them. Historically, the English language has had an advantage in this regard.

One concern for the future of AI models is the availability of open-source code, which all current models rely upon. As competition intensifies, companies may become less willing to share their code, which could impede further innovation by startups.

The featured image is from ceps.eu