Clément Delangue is, with Julien Chaumond and Thomas Wolfe, one of the three French co-founders of Hugging Face, a leading artificial intelligence (AI) start-up. Valued at €4.2 billion after raising nearly €450 million since its creation in 2016, this American company is known as the reference platform where developers and companies publish tools and models for making AI in open source, i.e. freely accessible and modifiable.
Read also | Article reserved for our subscribers Artificial Intelligence: In Paris, “Something is happening right now”
Add to your selections
Living between Miami, New York and Paris, where half of Hugging Face’s 220 employees are based, its CEO defends the design of an AI “open and decentralized”, against the domination of digital giants like OpenAI and its partner Microsoft or Google.
Hugging Face wants to “democratize good AI,” what does that mean?
An AI is ethical when it is aligned with the values of the individuals who participate in the project. At Hugging Face, our values are openness, transparency, and inclusiveness. We are therefore thinking about ways to make the technology as open as possible, to allow companies to share their AI models and the datasets and content on which they have been trained. We also seek to prevent one company from controlling the sector. We would like everyone to be able to create their own AI: companies, but also associations and NGOs, public structures, regulators, journalists, etc.
Read also | Article reserved for our subscribers “Artificial intelligence is a bubble: there is a gap between the costs, which are very high, and the potential revenues”
Add to your selections
Powerful players like Google or OpenAI offer large closed AI models, why is this a problem?
There are several complementary approaches. When you want to make a search engine, like ChatGPT, it’s good to use a very large AI model. But when you want to create a chatbot – a chatbot – customer service for your banking application, you don’t necessarily need an AI that talks about the meaning of life, but rather an AI that answers you quickly. A smaller and more specialized model is faster and allows you to spend less money and electricity, and therefore also emit less CO2.
Read also our survey (2017) | Article reserved for our subscribers Behind AI, the surge of data centers
Add to your selections
Furthermore, it is important that all people who want to build AI systems have access to the necessary resources, just as anyone can write computer code to create software. This is a more desirable world than one where a few very large companies keep their secrets and their AI models, and control access to this technology.
You have 80.28% of this article left to read. The rest is reserved for subscribers.
Related News :