What Are The Challenges To AI Adoption In Cloud Computing? 9 Experts Share Their Insights

  • 30 September 2019
  • Sam Mire

We are resistant to change. Because we derive comfort from familiar habits, the introduction of new products and approaches to any marketplace can be a tough sell. But when an innovation, in this case artificial intelligence, provides convenience and utility at a reasonable cost, you can almost bank on widespread adoption given enough time. But first, the marketplace has to be convinced that the benefits are worth the effort and cost of implementing a new mode of doing things.

Trends indicate that those benefits are worth, at the very least, some investment. The extent of the challenges facing AI adoption will go far towards determining the technology's ultimate role in cloud computing's future.

These industry insiders are well-aware of the challenges that face further AI adoption in cloud computing. Here's what they see as the most towering obstacles that lie in front of AI practitioners in the cloud computing space:

1. Brian Ray, Managing Director of Machine Learning for Maven Wave

“The #1 challenge is integration. Initial challenge is the movement of data and technologies to the cloud in the first place to allow for AI. Once it becomes available, and the technology is proven, there are complexities around the maintenance of those systems, the data, and the ability to adapt to an ever-changing environment. Cloud providers are inventing rapidly to lift the burden of these tasks.”

2. Christine Livingston, AI Chief Strategist at Perficient

There are concerns about security and privacy in leveraging cloud-based services. A lot of the data that organizations are trying to model is core to their business and includes sensitive personal, financial, health and other identifiable information. 

What’s important to understand is that a lot of deployments are specific to a particular industry and will have to meet respective certification requirements. Often times, this concern is satisfied by a vendor’s capability to create a complex and secure data center. Most of the widely reported data breaches are hackers getting into a single data system and not breaching a major cloud vendor’s system. As the accessibility and deployment of AI expands, concerns surrounding security and privacy will continue to breakdown.”

3. Bret Greenstein, Vice President and Global Head of Artificial Intelligence at Cognizant

Bret Greenstein“Cost, performance, workload management, and data access. These are all challenges that data scientists face as they do their work in the cloud. Ultimately, this is an optimization problem. In theory, the cost and speed should be fastest just running data science experiments on your own dedicated hardware. But, some problems are best solved with many VM’s, all spun up for very short times in parallel. We are already using AI to help optimize where to run workloads to balance cost, performance, and availability.”

4. Jeff Looman, Vice President of Engineering at FileShadow

“Considering the controversial nature of certain AI applications, the challenge can be stated in the old phrase “just because you can, does not mean you should.” AI raises many ethical questions. The challenge is determining what we should be doing, how should it be applied and whether or not these efforts make lives better or end up creating more grief.”

5. Steven Mih, CEO of Alluxio

“Moving AI workloads to the cloud is one of the biggest challenges we see companies facing. Many have data stored across silos, whether it’s in the cloud, in HDFS clusters, on-premises. Training models have more data than ever before, but that data is increasingly distributed and the number of queries is growing fast, putting more load on systems.

Hybrid latency prevents companies from running AI workloads in the cloud with data on-prem. So, most copy data into the cloud and maintain that duplicate data. And companies with compliance/data sovereignty requirements may even prevent organizations from copying that data. All of this means it is challenging to make both on-prem HDFS AI data accessible and high performing.”

6. Nima Negahban, CTO of Kinetica

“Data. AI in the cloud operates best with large quantities of data, and modern firms are certainly not lacking in this resource. However, the problem comes with ensuring data is clear and accessible so that AI can deliver value. When data is unstructured, incomplete, or siloed, AI solutions fall short and firms don’t see the true benefits of AI adoption.”

7. Michael Harrison, Managing Partner at Winterberry Group

“Inadequate data or data quality is the number one challenge to successful AI adoption. Marketers need to have their data strategy and infrastructure in place before implementing AI. Technology is not a solution for poor data or data quality.  Sparse data could cause a lack of meaningful insight or even worse, incorrect insight. Marketers are still struggling with the customer identity issue, which if AI is applied, will lead to false customer experience recommendations.”

8. Rob Clyde, Board Director at ISACA 

“Privacy, security and ethics are the biggest challenges – just because we can do something doesn't mean we should. For example, facial recognition powered by AI in the cloud is becoming more common place. However, when a person can be identified by their face anywhere, anytime, perhaps by anyone, everything we do could potentially be known and made public. In countries with totalitarian governments, this leads to a chilling and even dangerous situation. How do organizations ensure that the AI solutions they build and use meet appropriate security, privacy and regulatory requirements? How would we know if a machine learning algorithm has been maliciously or fraudulently trained by someone?”

9. Carl Hasselskog, co-founder and CEO of Degoo

“The challenge with AI lies in its implementation. Many companies understand the need to integrate the technology to benefit users, but they just as well do so from their own vantage point and neglect their own customers’ priorities in the process. The technology in Apple’s voice assistant Siri, for example, is underused compared to Amazon’s Alexa because users feel awkward talking to themselves in public, and would rather use the voice technology in the privacy of their own homes. Companies need to empathize with their users when implementing and ask “will this AI component make their lives easier?”

Have expert insights to add to this article?

Share your feedback and we'll consider adding it to the piece!


About Sam Mire

Sam is a Market Research Analyst at Disruptor Daily. He's a trained journalist with experience in the field of disruptive technology. He’s versed in the impact that blockchain technology is having on industries of today, from healthcare to cannabis. He’s written extensively on the individuals and companies shaping the future of tech, working directly with many of them to advance their vision. Sam is known for writing work that brings value to industry professionals and the generally curious – as well as an occasional smile to the face.