Why AI needs the cloud
This post is a response to a question. Before we get to that, I’ll provide some context. You may have heard about our new AI lab opening soon in Texas. The rumors are true. It’s the product of a partnership between the cloud organizations at Atos and Google. You may also have noticed that my most recent posts have stressed the role of data in AI and ML. Take this one for example: AI and ML without data? What are you going to do?
Now to the question. Someone asked me recently, “Why, if it’s all about the data, does AI sit in the cloud organizations of both companies and not their data analytics organizations?” Perhaps I’ve been remiss, at least in my blogs, by emphasizing the data and not the cloud.
So, why is our cloud organization the face of our AI program? Primarily because cloud is the delivery model. There’s no way around that, whether it’s private, public or hybrid cloud. That’s the first part.
Second, considering the computational horsepower required to drive the AI and ML models and their supporting IT, it can and does require a pretty significant technology underpinning to make it work. You can’t just run them on a small execution server. The larger the data sets get, the more complex the use cases get, the more compute is required — especially if you want to do AI in real time.
If you’re an enterprise using AI, you’re either getting those capabilities through the cloud or executing them on the cloud, or both.
I think that’s the evolution you’ll see when it’s over. AI fits nicely with cloud because cloud is meant to be on demand, at scale and global — the same basic characteristics you have for AI and ML platforms. So, there’s synergy between cloud as a platform and as the primary delivery vehicle for the technology and services required to do AI and ML.
The other reason cloud makes sense is because of the data that’s core to AI and ML. When we bring up data, we also have data sovereignty to deal with, meaning the European Union’s General Data Protection Regulation (GDPR). If you’re going to deploy cloud correctly globally, you’re going to do things like geofencing to control where the data resides.
A simple example is Amazon Web Services (AWS). The EU does not want data instantiated at AWS in a North American region, where their perception is the PATRIOT act can get access to that data. However, if the data stays in, say, the Frankfurt region in Germany, then the EU controls what happens to those physical servers and the government of the United States can’t touch the data.
With the cloud, I can use the same sets of AI technologies that I may have built anywhere in the world. I may use the same algorithms on data I have residing in the U.S. on an Amazon server, say, in the Virginia region. But if I have the data in Frankfurt, it’s geofenced and can’t leave that region.
The AI-cloud connection
That’s part of why cloud, AI and ML sit together in most of the pureplay cloud providers. Cloud is important because it’s global. And most AI and ML platforms that we deploy are for global companies, to help them understand, for example, what their data tells them about the customer base, or how to improve efficiencies.
If you want to explore AI for your organization, the cloud organizations of Atos and Google can help. More information about our new AI lab is available here.