AI is Coming, Prompting New IT Security Concerns
Intelligent, often autonomous, systems are making headway in the datacenter as developers seek to offload manual processes and move beyond traditional approaches like prescriptive IT automation as they struggle to keep up.
That’s the conclusion of a vendor-backed survey of IT executives about the adoption of intelligent machines and systems based on new AI approaches and other expert systems. Still, the survey sponsored by IT management software specialist Ipswitch notes that early adopters of automated systems are struggling to gauge security and access risks associated with handing the keys to bots and other electronic assistants.
“IT decision makers recognize that, while a force for good, these technologies also expose the enterprise to new internal and external risk vectors,” Tony Lock, an analyst with survey author Freeform Dynamics, noted in a statement. “As the pace of adoption increases, there will be no escaping the impact of intelligent systems on the enterprise, regardless of whether or not organizations directly invest in such technologies.”
The IT automation survey released this week also found that hyper-scale computing, advanced algorithms and huge datasets have helped make AI systems a “mainstream reality” in datacenters as well as the cloud. “The inclusion of advanced processing and machine learning [capabilities] into more familiar applications is then driving progress from the other direction,” survey authors noted.
There is little debate that the pace of IT operations is quickening and that DevOps teams have their hands full trying to keep up. In response, the survey found that the vast majority of those polled (91 percent) agreed that technology advances allow more IT operations to be automated.
While the early focus has been on handing off mundane, non-mission-critical tasks to machine learning and other automation tools, a shift to more advanced systems is being driven by, for instance, Internet of Things deployments. IoT platforms and development ranked highest on the “to-do” list of survey respondents. Other emerging technologies cited by executives included complex event processing technology and “rules-based, process-automation and workflow engines.”
IT professionals “should approach intelligent machines with their eyes wide open,” acknowledged Jeff Loeb, Ipswitch’s chief marketing officer. “As network managers grow more confident with intelligent systems, they will become increasingly willing to tackle more complex applications.”
Previous industry surveys have documented trust issues among middle managers about the rise of machine intelligence. In a report released earlier this year on the promise and pitfalls of using intelligence machines in the workplace, market analyst Accenture Strategy argued the technology could help free up managers’ time to focus on strategic initiatives. Nevertheless, the study found several hurdles for widespread adoption, including a “trust gap within the managerial ranks.”
By contrast, the Ipswitch sponsored survey found that 76 percent of IT executives polled said machine intelligence could reduce drudgery from IT operations while only 32 percent were worried that automation would displace them.
Rather, concerns about deployment of automation tools focused on security issues, especially unforeseen consequences like unauthorized API and file access. “Security and access measures already in place may not have been designed to deal with the new and different types of automated activity that will rapidly become the norm, let alone the likely increase in traffic and events that will need to be handled,” the survey warned.