As a bona fide information technology professional, you have been inundated recently with the idea that artificial intelligence and machine learning are going to change the way your enterprise does business. Time and time again you have been told that, coupled with big data and IoT, AI is going to transform the enterprise network and how it is managed and that an IT pro’s life will never be the same.
Even if you cut out the hyperbole, however, much of that scenario does remain intact–to keep up with the demands of modern enterprise-generated data, more devices will be required to have some form of functional, self-contained AI. In fact, companies like Microsoft are counting on it. More AI devices creating more data means more demand for cloud resources, which means more profits for the companies providing the services.
Of course, before such a modern AI-based enterprise network can become reality, a great deal of research and development must be completed. But testing AI systems requires datasets and metrics with known parameters–something that does not exist in great numbers. This is one of the reasons why Microsoft has made many of the AI research datasets it has created publicly available.
SEE: Research: Companies lack skills to implement and support AI and machine learning (Tech Pro Research)
Share the wealth
According to a November 2017 AI Blog post, the Microsoft Research Maluuba team published its FigureQA dataset, including metrics and other tools for testing AI systems, to provide a codified way for everyone from academic researchers to industry experts to test their systems, compare their work, and learn from each other.
While in many ways, this sort of research sharing goes against typical industry practice, it is practically essential if AI is going to become a dominant feature of the modern enterprise network in the next few years. It is too inefficient for each major player in the information technology industry to do its own independent AI research. Cooperation and collaboration offer the only sensible way to advance the research in a timely manner.
When it comes to AI development and testing, it is vital that the results generated by an AI system be compared with known, predetermined metrics to check for accuracy and consistency. The more known datasets and metrics to test an AI against. the surer researchers can be about an AI once it is released as an application in an enterprise network.
SEE: AI will eliminate 1.8M jobs but create 2.3M by 2020, claims Gartner (TechRepublic)
Before AI can become a dependable, reliable part of the enterprise network, it must first be thoroughly researched and tested. No individual company, even one with extensive resources like Microsoft, can do all this testing alone. AI research requires cooperation and collaboration among all the major industry players.
By sharing AI datasets, metrics, and tools with the public, Microsoft is looking to create a research environment that grows the overall AI market at a faster pace than it could accomplish on its own. A faster developing market, complete with a quicker rate of adoption, means a larger pie from which to cut its own market share.
In general, this is an understandable and reasonable business strategy on Microsoft’s part. However, it remains to be seen what the widespread implementation of AI in the enterprise will mean for information technology professionals. While AI may be good for enterprise business in the long run, it is sure to be disruptive and challenging during deployment. IT pros should be prepared for more pain before realizing any gain.
- Here are the top 3 benefits and barriers to AI adoption (TechRepublic)
- Special report: How to implement AI and machine learning (free TechRepublic PDF)
- The 6 most in-demand AI jobs, and how to get them (TechRepublic)
- The greatest risk with AI is not moving fast enough to deploy it: Microsoft (ZDNet)
- Microsoft reveals Azure IoT Edge: Putting AI at the furthest reaches of your network (TechRepublic)
- How Microsoft plans to turn Azure into an ‘AI cloud’ (ZDNet)
- Don’t be alarmed, but you’re probably using the term AI wrong (ZDNet)
Are you ready for AI devices? Is your enterprise ready for AI? Share your thoughts and opinions with your peers at TechRepublic in the discussion thread below.