Monday was a big day for announcements from tech giant Microsoft, unveiling new guiding principles for AI governance and a multi-year deal with Mistral AI.
Tech behemoth Microsoft has unveiled a new set of guiding principles on how it will govern its artificial intelligence (AI) infrastructure, effectively further opening up access to its technology to developers.
The announcement came at the Mobile World Congress tech fair in Barcelona on Monday where AI is a key theme of this years event.
One of the key planks of its newly-published "AI Access Principles" is the democratisation of AI through the companys open source models.
The company said it plans to do this by expanding access to its cloud computing AI infrastructure.
Speaking to Euronews Next in Barcelona, Brad Smith, Microsofts vice chair and president, also said the company wanted to make its AI models and development tools more widely available to developers around the world, allowing countries to build their own AI economies.
"I think it's extremely important because we're investing enormous amounts of money, frankly, more than any government on the planet, to build out the AI data centres so that in every country people can use this technology," Smith said.
"They can create their AI software, their applications, they can use them for companies, for consumer services and the like".
The "AI Access Principles" underscore the company's commitment to open source models. Open source means that the source code is available to everyone in the public domain to use, modify, and distribute.
"Fundamentally, it [the principles] says we are not just building this for ourselves. We are making it accessible for companies around the world to use so that they can invest in their own AI inventions," Smith told Euronews Next.
"Second, we have a set of principles. It's very important, I think, that we treat people fairly. Yes, that as they use this technology, they understand how we're making available the building blocks so they know it, they can use it," he added.
"We're not going to take the data that they're developing for themselves and access it to compete against them. We're not going to try to require them to reach consumers or their customers only through an app store where we exact control".
The announcement of its AI governance guidelines comes as the Big Tech company struck a deal with Mistral AI, the French company revealed on Monday, signalling Microsofts intent to branch out in the burgeoning AI market beyond its current involvement with OpenAI.
Microsoft has already heavily invested in OpenAI, the creator of wildly popular AI chatbot ChatGPT. Its $13 billion (11.9 billion) investment, however, is currently under review by regulators in the EU, the UK and the US.
Widely cited as a growing rival for OpenAI, 10-month-old Mistral reached unicorn status in December after being valued at more than 2 billion, far surpassing the 1 billion threshold to be considered one.
The new multi-year partnership will see Microsoft giving Mistral access to its Azure cloud platform to help bring its large language model (LLM) called Mistral Large.
LLMs are AI programmes that recogise and generate text and are commonly used to power generative AI like chatbots.
"Their [Mistral's] commitment to fostering the open-source community and achieving exceptional performance aligns harmoniously with Microsofts commitment to develop trustworthy, scalable, and responsible AI solutions," Eric Boyd, Corporate Vice President, Azure AI Platform at Microsoft, wrote in a blog post.
The move is in keeping with Microsoft's commitment to open up its cloud-based AI infrastructure.
In the past week, as well as its partnership with Mistral AI, Microsoft has committed to investing billions of euros over two years in its AI infrastructure in Europe, including 1.9 billion in Spain and 3.2 billion in Germany.
See the original post here: