Following Microsoft's investment in and partnership with French large-scale language modeling startup Mistral AI, the company is working hard to shake off its image of stifling competition through its deep partnership with (and financial stake in) OpenAI. I keep trying. Today, the company launched a new framework it calls the “AI Access Principles.” This is an 11-point plan that Microsoft says will “define how AI data center infrastructure and other critical AI assets around the world will be operated.”
These points cover areas such as building and operating app stores that allow companies to choose from a variety of LLMs and other AI products, and efforts to keep training models from including proprietary data. It also includes a promise to allow customers to change their cloud provider and services within the cloud if they choose. We also detail our focus on building cybersecurity around AI services. Pay attention to building data centers and other infrastructure in an environmentally responsible manner. and investment in education.
Brad Smith, President and Vice Chairman of Microsoft, announced the framework today at Mobile World Congress in Barcelona. The implication here is that Microsoft is open to dialogue and dialogue with stakeholders, but the irony is that Smith announced this news in his keynote, leaving no room for follow-up questions. There was no.
The announcement comes as Microsoft faces increased regulatory scrutiny of its $13 billion investment in OpenAI. OpenAI has given it a 49% stake in the startup, which is currently at the forefront of global generative AI services. Europe's competition watchdog announced in January that it was assessing whether the investment fell under antitrust laws.
The plan is for third parties to use Microsoft's platform and services to collaborate with carriers participating in MWC, as well as businesses and organizations around the world, which the company hopes to develop over the next few years. We are specifically targeting how to develop AI products, a key business area and enterprise service in the world. A much wider industry.
“If they are training a model on our infrastructure and deploying it on our infrastructure, we recognize that their data is their data. will not access it or use it to compete with companies that rely on our infrastructure,” Smith said.
To be clear, these AI Access Principles are not binding rules for Microsoft. Additionally, no details are provided as to how commitments will be verified or tracked. But they serve a purpose in anticipation of that. If a formal regulatory investigation is conducted; These are likely to be used by companies to claim that they are making active efforts to ensure competition in the market.
“In fact, as of today, we have approximately 1,600 models running in our data centers, 1,500 of which are open source models,” Smith said on stage today. Both large and small companies. ”
On the other hand, this public pronouncement of principles provides a public declaration that the public, Microsoft's competitors, and strict regulators can use as a reference point should they decide that Microsoft is not up to the task. It becomes.
