Artificial intelligence is firmly wedged into the Gartner hype cycle, but that does not mean it is vaporware or that its impact on the enterprise will be less than stunning. It does mean that some expectations need to be brought back to earth.
According to Sarah Thomas, director of Women in Comms, AI is likely to be baked into nearly every new software product and service by 2020, but at the moment the rush to pin the AI moniker on existing systems is causing confusion. This is not altogether uncommon, of course. The enterprise saw much the same thing with virtualization, the cloud and, lately, the Internet of Things. But it does mean that both users and developers need to take a step back to assess what AI really means and how it should be employed to the greatest benefit to everyone.
This effort won’t be helped if every technology provider starts proclaiming itself an AI company, as appears to be happening now, says Gartner’s Jim Hare. In a paper highlighted by The UK Register, Hare notes that this kind of term washing is always self-defeating because it leads to empty promises that result in disillusionment and depressed uptake in the near term. The most common form of this practice at the moment is taking a rules-based machine learning and analytics platform and calling it AI, even though it has no capacity for self-learning or broad situational awareness.
Fortunately, it is becoming easier for the enterprise to check out all manner of smart technologies now that the leading cloud providers are taking up the cause. Jake Bennet, CTO of Seattle tech consulting firm POP, says it is now possible to watch a quick tutorial, spin up a 10-node cluster and test out multiple intelligent technologies, all for a few hundred dollars. This is, in fact, turning into the newest front in the cloud wars as giants like Google, Amazon and Microsoft try to corner the market on AI-infused services. The hope is that, before long, the gap between AI users and non-users will be so great that organizations will be scrambling to align themselves with providers offering the most advanced capabilities.
Private and hybrid clouds are also gaining quick on-ramps to AI-powered services. A company called EasyStack recently introduced what it calls the first OpenStack-based AI cloud platform that supports GPUs and FPGAs. The ESCloud works in conjunction with a new Linux container system and both PaaS and CaaS implementations to allow organizations to build and deploy high-performance AI architectures in record time. There is also an Express version aimed at small-and-medium businesses that allows them to maintain parity with larger competitors.
The precise definition of artificial intelligence will probably be a moving target for some time given the fact that no algorithm is likely to achieve true intelligence in a biological sense. That gives marketers all the license they need to define the term to suit their own purposes – barring, of course, a hard and fast legal definition, as found in whiskey and other products.
But this is all the more reason for the enterprise to forget about the labels and focus on the processes. In the end, it doesn’t matter what it’s called as long as it gets the job done in a better way than what you have now.
Arthur Cole writes about infrastructure for IT Business Edge. Cole has been covering the high-tech media and computing industries for more than 20 years, having served as editor of TV Technology, Video Technology News, Internet News and Multimedia Weekly. His contributions have appeared in Communications Today and Enterprise Networking Planet and as web content for numerous high-tech clients like TwinStrata and Carpathia. Follow Art on Twitter @acole602.