What you need to know
- Microsoft has faced increasing scrutiny toward investments with AI tech start-ups like OpenAI and Mistral AI as concerns mount that these partnerships could block competition in the nascent AI market.
- To curb concerns, Microsoft president and vice chair Brad Smith has announced Microsoft's new AI Access Principals for "the AI economy" during the Mobile World Congress event in Barcelona.
- The plan includes investments in AI data centers, skilling programs, efforts to combat deepfakes, and partnerships with global AI start-ups.
Microsoft Vice Chair and President, Brad Smith, took to the Mobile World Congress event in Barcelona to outline the tech giant's commitment to accessible AI technology. Microsoft's AI Access Principals feature an 11-point plan aimed at providing access and support for developers to AI models and tools while committing Microsoft to "providing broad technology access to empower organizations and individuals around the world to develop and use AI in ways that will serve the public good."
While debates rage over the ethics of generative AI, major tech companies like Microsoft have not held back on investing in the technology. Microsoft has thrown large sums of money behind language modeling startups like OpenAI, of which Microsoft owns a 49% stake. These investments have begun to generate some concern, both from competitors and antitrust regulators, that Microsoft is positioned to squash competition in the burgeoning AI marketplace.
AI Access Tenets
In the last two weeks, Microsoft has invested more than $5.5 billion in AI data centers in Berlin and Madrid and penned a multi-year partnership along with a $16 million investment in French tech startup, Mistral AI. Microsoft also announced plans for AI skilling programs and launched a Tech Accord of 20 companies dedicated to combating deepfakes in elections through public awareness and education programs. Microsoft recently previewed a new tool using AI to protect consumers from scam calls and new technology to further advance the Open Gateway Initiative in a bid to improve developer access to AI services.
Smith's framework for accessible AI technology begins with Microsoft's core tenets as a provider of infrastructure and platforms for AI access, including the company's responsibility toward innovation and competition. This particular tenet is an attempt to assuage the global regulators who have taken issue with Microsoft's investment activities.
In the blog post detailing the AI access principles, Smith points to the limited availability of GPUs as the bottleneck for access to AI technology. In an acknowledgment that the GPUs that power AI data centers are primarily produced by one company, Smith declares that Microsoft is required to "live in the land of partnerships" to succeed.
Smith also acknowledges Microsoft's responsibilities and obligations under the rule of law as one of its core tenets. New laws and regulations regarding the use of AI are cropping up in The European Union, the United States, and more countries around the globe. "We take these obligations seriously," says Smith of the new legal challenges.
As part of the new AI access tenets, Microsoft has declared a commitment to furthering its partnerships by being "proactive and constructive." These partnerships are an effort on the part of the corporation to get ahead of accusations that Microsoft could potentially squash competition by picking and choosing the startups to back.
A similar argument was made by the UK's regulatory authorities during Microsoft's merger with game publisher Activision Blizzard King. To calm regulators' concerns, Microsoft was ultimately forced to sign an agreement giving cloud streaming rights to ABK's catalog of games to Ubisoft. When it comes to AI, Microsoft has attempted to skirt this concern by providing seed funding to an array of startups around the world.
AI access principles
In addition to commitments to core tenets and partnerships to further improve access to AI, Microsoft laid out an 11-point plan to provide access and support for developers in an effort to improve models and applications that are created using AI innovations. According to Microsoft, this plan will foster competition by enabling access to software to developers globally.
Microsoft's plan can be summed up as so:
- Grow chip capacity to further expand Microsoft's cloud computing AI infrastructure.
- Make AI models and development tools broadly available to developers globally.
- Make public APIs available to developers to use and access AI models hosted on Microsoft Azure.
- Support a common public API to further support software developers.
- Provide the freedom and choice for developers to distribute their AI models, tools, and applications on Microsoft Azure via the Azure marketplace or directly to consumers.
- Ensure non-public information or data is not used to train, build, deploy, or otherwise use developers' AI models in Microsoft Azure to compete with said models.
- Enable customers to switch from Microsoft Azure to another cloud provider with easy export and data transfer tools.
- Support the physical and cybersecurity needs of all AI models and applications that utilize Microsoft's AI data centers.
- Apply a strong 'Responsible AI Standard' to keep people at the center of AI design that respects values such as "fairness, reliability, safety, privacy and security, inclusiveness, transparency, and accountability."
- Invest in initiatives to improve the spread of AI skilling globally.
- Manage data centers in a way that is environmentally sensitive and use AI to advance sustainability needs.
Apart from antitrust concerns, the ethics of powering and sustaining the data centers required for AI technology has also been a source of contention. Microsoft's leadership has insisted that the company is taking steps to ensure that the power draw to run AI models and data centers is sustainable, despite reports that the actual consumption rate for Generative AI could power a small country. The company continues to look into alternative power sources to keep Azure data centers up and running, including considering nuclear power as a solution.
Thoughts on AI economy
Regardless of whether you love or hate AI, major tech players like Microsoft and Google and concerned market regulators alike see a future where the technology has carved out its own sector of the economy. The legislative process that is necessary to protect consumers and civilians from being taken advantage of by corporations in these scenarios often lags way behind the advancing technology.
While it still amounts to being a band-aid on a proverbial axe wound, Microsoft's preemptive efforts to commit to standards of fairness and competition is a step in the right direction. What matters most, however, is that the "powers-that-be" in society put the checks and balances in place to ensure that those commitments are carried out as promised.
