Microsoft to Invest 10 Billion in OpenAI
Microsoft and OpenAI are teaming up to build the future of AI
Created on January 25|Last edited on January 25
Comment
Following investments in 2019 and 2021, Microsoft has extended their partnership with OpenAI, and will be investing a rumored 10 billion dollars in them. The deal also includes agreements allowing OpenAI to leverage the continuously increasing scale of Microsoft's Azure cloud platform.
Microsoft will benefit from this deal through agreements allowing them to use many of OpenAI’s technologies in their enterprise cloud services. The two companies also cite their shared values as a key reason for the partnership.
OpenAI has invested heavily in its “Alignment Research” which focuses on building intelligent systems that are “helpful, truthful, and safe.” They cite the main pillars of their alignment research as
- Training AI systems to assist human evaluation
- Training AI systems to do alignment research
Although these principles are thoughtful and well-intended, there are some drawbacks that OpenAI acknowledges.
If humans with poor intentions can align AI systems with their own motivations, it could lead to undesirable outcomes. OpenAI plans to stay on the cutting edge of the technology to adapt quickly as these challenges arise, as they see technical aptitude as the only solution against bad actors.
In their third pillar, they also mention that its unlikely human researchers will be able to keep up with the pace of the alignment challenges, so they plan to develop AI systems that are capable of carrying out alignment research automatically, as they see this as the only scalable solution to the non-linear pace of the alignment challenges.
These principles as well as OpenAI’s capped for-profit status, make the company well-suited for building out the future of AI safely; however, it's unclear if they will be enough to defend against the investible adversaries that may arise, and it will be interesting to see how other large organizations like Palantir and Google respond to the trend of prioritizing alignment in AI research.
References
Add a comment