Buying AI tools is easy. Building the infrastructure to support them is not as easy.
According to John Hines, Senior Director for Enterprise Business at Lumen Technologies, say that many companies are charging headfirst into AI without understanding the operational, security and networking demands sitting underneath it all.
“We’re seeing the adoption of AI across Australian enterprises is rising pretty rapidly, but it’s outpacing technical preparedness in many cases,” Hines said during a recent interview.
“Companies aren’t struggling with AI itself, they’re actually struggling with the infrastructure that wasn’t designed for the scale of AI workloads.”
Hines said traditional enterprise networks were built around predictable, human driven activity, not massive real time AI data flows spread across cloud environments.
“AI workloads require low latency, high throughput, consistent reliability,” he said. “Without that, even the best models are going to underperform.”
Realistically, the warning comes as executives scramble to prove they are not ‘falling behind’ competitors in the AI olympics. But according to Hines, that pressure is creating a dangerous “all you can eat” mentality inside some businesses.
“We’ve seen organisations that six different parts of the organisation are running six separate pilots in isolation of each other,” he said. “They’re lacking from an efficiency perspective, from a learning perspective.”
While companies continue pouring money into AI projects, many still cannot explain what return they are actually getting.
“There’s a huge amount of hype around AI,” Hines said. “Everyone’s kind of jumping in on it. But realistically, what is the cost of it and what is the cost benefit?”
Some organisations have already discovered the hard way that replacing humans too quickly can backfire and there are now companies crawling back to pick up the pieces.
“There have been some organisations who have removed people out of their organisations only to figure out that they actually need those humans and have rehired them back,” Hines said.
Hines believes the AI frenzy will eventually cool off as businesses realise the technology is not a ‘silver bullet’.
“Perhaps AI is not the answer to everything,” he said. “Perhaps you do need some humans in place to do certain things. Perhaps it wasn’t going to be the silver bullet that’s going to save an organisation a fortune and enable it to operate with two people.”
Still, the technology is moving fast, the pace of decisions and so are the risks.
Hines warned companies are underestimating how exposed they could become if AI systems are not governed properly. One of the biggest threats is not just hackers attacking AI systems externally, but organisations poisoning themselves internally through poor controls and unrestricted access.
“If you don’t have the right guardrails in place and other employees start to query that model, I don’t think it’s too difficult to figure out very quickly what everyone’s getting paid,” he said, referring to internal AI systems trained on payroll or HR data.
He also pointed to growing fears around manipulated AI systems and chatbot abuse, referencing examples where researchers convinced AI-powered sales systems to generate absurd offers.
“Unless you’ve got those guardrails in place, i’s quite easy to manipulate a chatbot to give you what you’re asking for,” Hines added.
Mr Hines said companies that take a disciplined approach, alongside proper governance, infrastructure upgrades and clear business outcomes, are likely to emerge ahead of the pack.
“I think those who will be successful are those who have properly planned it out,” he said. “AI is just not a magical thing you throw out there and it’ll solve everything.”
