The global AI market is expected to grow from $21.46 billion in 2018 to $190.61 billion by 2025. No doubt, it’s one of the fastest-growing markets in the world today. Pretty impressive, huh?
On the other hand, however, the public cloud industry stood at $182.4 billion in 2018 and is projected to grow 17.5% in 2019 to total $214.3 billion. Unlike AI, the cloud industry has already trodden the path from hype to broad adoption, and become a different beast altogether.
With the cloud industry propelling this sort of growth, could the forecasts about AI adoption ultimately prove to be conservative?
Or, perhaps it’s time we realised that the AI and cloud computing industries are not mutually exclusive.
What is artificial intelligence? Defining it in business — a CTO guide
In the past decade, cloud computing has gone from been seen as a cost-saving way to store data and applications to becoming an integral part of advancing AI and other cognitive capabilities in the enterprise.
In fact, a recent survey by Deloitte found that 49% of companies that have deployed AI today are using cloud-based services.
“Cloud adoption is motivating enterprises to undertake more proofs of concept in their firms with AI because it’s easier than ever before to get started,” said David Schatsky, managing director at Deloitte LLP.
According to Schatsky, this path is also becoming more attractive to enterprises as cloud providers continue developing AI offerings to business functions, without big upfront costs.
Integrating artificial intelligence: Best practice tips — a CTO guide
AI without the cloud is tough
The cloud is fundamental to the AI model in two ways: first, given that the data sets these companies are using would not be accessible if it was not for the cloud, and secondly, because only the cloud can enable businesses to cope with the phenomenal scale required by providing such data-intensive services to multiple clients at an affordable cost. Integrating AI into existing processes and workflows is another major challenge.
Separately, one of the biggest obstacles holding AI back from mass adoption is the shortage of people within enterprises with the skills to programme it. This means that while businesses may know how they want to use AI, they don’t have the means of building an application or algorithm to produce the results they crave.
The cloud changes this as it means that years of research and tools are available to developers tasked with creating AI solutions, even if they lack the expertise to build and train systems, or to manage data on their own.
What industries will AI impact the most in the next few years — a CTO guide
Is cloud-based AI scalable?
According to forecasts by Tractica, the technology market research firm, AI will account for as much as 50% of total public cloud services revenue by 2025.
Needless to say, the cloud giants aren’t sitting around twiddling their thumbs; they’re bringing competitive AI offerings to market at an agile rate.
“These include hardware optimised for machine learning, application programming interfaces that automate speech recognition and text analysis, productivity-boosting automated machine learning modelling tools, and AI development workflow platforms,” explained Schatsky.
For instance, Walgreens, the American pharmacy store chain, is planning to implement Microsoft’s Azure cloud-based AI platform to develop new health care delivery models, while The American Cancer Society uses Google’s machine learning cloud services for automating tissue image analysis.
But will this path lead to success?
“The various AI API’s that have been made available by the major cloud players are a great resource for companies to get started and try things out,” said Schatsky. “However, we find a lot of companies who want to be aggressive in their AI adoption don’t have the right foundations in place.
“If an enterprise wants its AI to go from a proof of concept stage to being enterprise-ready, they’ll need to get their data together and accessible in a central place. So, there’s a lot of preparatory data engineering and data wrangling work that needs to be done.”
Furthermore, if enterprises are deploying cloud-based AI solutions from multiple vendors it can all get a little complex, especially when it comes to efficiently weaving multiple different clouds together in a way that doesn’t hinder productivity or innovation.
The inherent technical challenges of managing several different platforms in a multi-cloud deployment — such as understanding exactly where data resides — has the potential to cause stress. They can also amplify already-existing business pressures, most notably responding to customer demands for service and innovation. This is especially true for those organisations that either lack the in-house skills or fail to find a partner that has the capabilities required to manage such a transformation.