Data

Influx of AI Products Brings New Benefits, Challenges

As adoption of AI increases and we move from a reactive to a predictive model, IT and data center pros need to carefully weigh risk and reward.

The number of products that include artificial intelligence and machine learning capabilities is rapidly increasing, and new technologies that make it possible to collect more data faster are becoming available. At the intersection of these opportunities lies risk, which means IT and data center pros need to strike a delicate balance to reap AI and ML rewards.    

Steve Zhu, VP of technology at Finn AI, noted that, despite the many AI-oriented products being introduced, it's still just the tip of the iceberg. The reality, he added, is that current applications of the technology barely scratch the surface of its full potential and capability.

“Integration of AI assistance in our lives is more than just about automation and optimization, but increasingly about how it can improve [well being], such as in the area of personal finance,” he said.

From the enterprise perspective, that means AI deployed across the internet of things, along with constant analytics and the massive amounts of data that comes with that, said Jane McConnell, a marketing executive at EON Reality.

But this also points to the need for new and additional security and storage.

“If the machines we work with on a daily basis can speak not just to us but with us, then the lines and outcomes of communication and throughput with those machines increases--so the need for security increases,” McConnell said. “That, in turn, means we could see a meteoric rise in the need for decentralized cloud storage.”

The more embedded AI becomes in our home and work lives, the more important security and privacy concerns will become, said Finn AI's Zhu. “Organizations asking consumers to embrace AI must also support it with trust, not only in data security, but additionally for the best interest of consumers,“ he said.

What Comes Next?

The emergence of specialized providers of “basic AI skills” will make AI development more efficient, Zhu said. “For instance, ‘best in class’ standards for AI-enabled voice, visual or audio technology will allow developers to efficiently leverage those capabilities into their product domain,” he said.

As AI applications like these become more ubiquitous and more reliable, there will be more familiarity and less hesitancy to bring them into the workplace, said Steve Stover, VP of product at Samanage.

“Users have to be confident in the technology’s predictive capabilities to make it work in enterprise IT, and bridging the divide between the average consumer and ‘smart’ services alleviates fear, driving up AI’s value,” he said.

There are areas in which AI is already making an impressive mark--for example, natural language processing and autonomous vehicles, said Moody. But, he added, AI that truly mimics human thought is still in its infancy.

Meanwhile, the data just keeps on piling up, boosted in part by the introduction of technology like 5G and 802.11AX that will accelerate the amount of data that AI and ML need to learn. This will speed up development of the tech itself, Moody said.

A new type of chip manufacturing--AI chips, also called Tensor Processing Units, or Neural Network Processors--will also boost the amount of data companies need to deal with, said Drew Farnsworth, a data center consultant and partner at Green Lane Design.

The chips can deal with huge amounts of data, Farnsworth said, but also require high-density data center space--another organizational challenge that comes with the push toward AI. Of course, huge amounts of data are necessary for effective AI and ML, and valuable for organizations, but all of that data requires storage capacity and IT and analytics staff who know how to use make use of it. Considering the existing AI staffing crunch, this will continue to be a significant challenge for any organization that sees a need for AI/ML but lacks skilled personnel.

Shift from Reactive to Predictive IT

When they aren’t planned properly, changes to IT infrastructure can cause more problems than they solve, Stover said. "We are still in the beginning stages of our shift from a reactive world to a predictive world,” he said. “As businesses continue to digitally transform and adopt new technology this year, IT leaders will need to better manage how these systems all fit together.”

For example, as use of the IoT increases, more businesses will use AI-driven models to predict the risk of a change before its implementation, Stover said.

We’ll really know when AI is no longer a buzzword and simply an expected part of products and services when we stop hearing about it so often, Zhu said. The technology’s presence will just become assumed.

Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish