How will the rise of artificial intelligence affect EMM, desktop virtualization, and EUC?

AI has a lot of momentum right now, so it’s worth paying attention so we’re not caught off guard.

Everyone is talking about artificial intelligence right now.

We expect the tech giants to be in on this (here are some examples regarding Google, AppleFacebook, and Amazon), but AI is very important to enterprise IT vendors like IBM and Microsoft, too. (See Satya Nadella's AI keynote from back at Ignite.)

(Also, I’m assuming it goes without saying, but just in case, we're not talking about general AI like sentient robots, rather we’re talking about more specific types. Over the last few years there have been tremendous breakthroughs in machine learning, involving deep learning and neural networks. Here’s a primer from the Andreessen Horowitz blog, a16z.)

Anyway, it’s time ask: How will the rise of artificial intelligence affect EMM, desktop virtualization, and EUC?

A few initial thoughts

As EUC folks, we’re concerned with deploying, managing, and securing different types of devices, apps, data, and identities. So to recast today’s question, we could ask: Does AI produce new types of devices, apps, data, and identities for us to deal with?

You could argue that AI will mostly affect the business logic within apps and devices themselves. So AI will mean more new apps, and refreshing or replacing old apps, but they’ll still be delivered with the same concepts we’ve been dealing with for years—they’ll be desktop, web, or mobile apps, or maybe within devices. The same goes for data. If AI is mostly about business logic within apps, and we already know how to deal with them, then we should know how to deal with their output, too, whether it’s documents, APIs, or dashboards.

Naturally, AI is also coming to the business logic of security and management products; so IT’s apps are getting it, too. For security, many use cases are emerging, such as products that use AI to spot things that look bad, whether it be malware, malicious users, hacked accounts, or insecure configurations. For management, imagine having AI to help a user get access to all the apps and data that they need—we could have all the settings and apps ready to provision right when they request it. AI is also becoming part of the core productivity apps and the devices that we use every day.

What about more problematic effects of AI? Well, there could be many, and of course we can’t predict them all. AI helps enable better voice-controlled user interfaces (Siri), so that’s a potential worry. (Does Siri have to know about enterprise DLP?) And a lot of AI execution will happen at the edge, in small devices, so we can’t just think of them as dumb clients, and instead we’ll have to worry about them even more. These are just two of many potential issues to ponder.

Why worry about AI?

Today I’ve just scratched the surface with a few initial thoughts, but overall I know that AI is something that we EUC folks need to pay attention to. Whatever it turns out to be, we at least want the adoption to be smoother than when mobile and cloud got us all worried about BYOD and Shadow IT. Also, looking again to what forward-looking organizations like Andreessen Horowitz are thinking, there’s a pretty good chance that AI is going to be the next big thing to follow mobile and cloud. So this is why we should start thinking about AI.

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

AI is already there in use on the virtualization and cloud space as I write this mail as I work on one such domain. The good set of AI algorithms linked to the data lakes on cloud will be of immense potential. Pattern analysis, reporting and de-duplication/optimization as just tip of the ice-berg.
But right now (within my limited knowledge) I sense only AI blocks which are customized for a specific purpose and nothing but limited compute models, contained on a specific platform.
 But just like a common platform for all other systems can increase the potential of using the technology, we need a common platform for developing on AI which is aimed specific in using on the cloud space. Once we have a common platform and workspace on all cloud spaces, then we can built AI algorithms which can span on multiple compute space across all cloud spaces. I tried one such using Apache Karaf based containers in which the AI algorithm was used as a spouts which are linked to each other via ssh as a ANN (Artificial Neural Network). The control node will spawn the containers as needed on a cloud space and link them as per the ANN design; and also pump in the AI algorithms and weights to the Karaf spouts. The monitor node will check the Learning algorithm and team this set with sample data and adjust the weights on the karaf nodes, via ssh. Since spawning of containers on the cloud is also automated and templates can be saved, we can spawn an ANN in a matter of minutes and start training them with a constant flow of data from a data lake, once the ANN is fully trained for a functionality. This is ANN based and hence almost real time computation happening across all containers and hence data processing speed will be high. Also the ANN layers can be scaled up and down as needed on cloud.

Here I am speaking of developing a brain on the cloud... the real AI stuff than locked down compute models.

(Fuzzy Brain Management tools…“Let your computer and applications work for you”) I can see the day where many of the software tools we use will have much more AI pre-configured out of the box. You will simply set behavior modes and activity level tolerance. Obviously some of this already exists. It is hard to see how this will radically change our office environments in the next few years. I think the computer OS in the movie “Her” is realistic when it shows how the future daily human computer relationship could be. The OS condenses and distills lots of information. Then is able verbally discuss the topic or task. The human speaks plainly to the computer about the results desired.