AI is an inescapable topic in any business or technology-linked conference in 2026, with sessions often presented by vendors crowing about the abilities of LLMs to transform businesses, up productivity and cut costs.
One of the areas arguably most exposed to the benefits and downsides of the hugely-hyped technology is security, with those dedicated to protecting enterprises seeing an increase in AI-aided attacks but also viewing the technology as a part of the defence.
So it was no shock to see security front and centre at conference Infosecurity Europe 2026 held earlier this month in London, with almost half of the nine keynotes on the opening day featuring AI in the title and the topic frequently mentioned across a large proportion of the other sessions.
During a panel comprising cybersecurity professionals working for McDonald’s, Live Nation Entertainment, IHG Hotels and Resorts, and international law company RPC, there was optimism about the prospects of AI as a defensive tool, though its limitations and potential threat from the technology when used by criminals was also raised.
On the attack.
Spencer Scott, head of information security at RPC (pictured below, far-left), said as with other sectors the legal industry is experiencing a “big uptick in deep fake voice cloning, where you have people impersonating C-suite” taking the form of spoofed executives joining voice calls and phishing through messages.
McDonald’s senior analyst cyber defence Ellie Hallam (pictured, far-right) argued this type of AI-driven trend is especially problematic for organisations reliant on distributed workforces.
“I’d argue if you have organisations reliant on remote workers you’re far more likely to fall victim to that because [attackers] are so good at using AI voice cloning,” she added, noting in an earlier session at the event it was revealed “you now need three seconds of voice to clone someone’s voice”.
Citing current reports on industry trends, Hallam highlighted a consensus AI is accelerating attacks, so “it’s just much more difficult for us as defenders to catch things when we need to, before they’ve become significant risks to the business”.
In terms of how attackers are structuring assaults, she indicated automation for selected tasks is more common than blanket AI being present across every element.
Boring jobs.
A regular selling point of AI across a lot of industries is the ability to take away repetitive tasks which often involve scything through huge volumes of data. These roles are unlikely to be enjoyed by staff or viewed as a good use of an experienced professional’s time.
For these labour-intensive tasks the panel was united in agreeing the technology offers plenty, though the need for human oversight was repeatedly stated.
However, not everyone was convinced by a lot of the vendor claims around use of the technology.
Live Nation Entertainment senior cyber threat intelligence analyst Alice Smith (pictured above, second from left) described AI as a “buzzword” and said she is not the “biggest fan”.
“It’s so hard to sift through all of the noise, all of the information” about the technology, she added, noting it is “always going to be a tool that’s going to be used to make things quicker and to make things more productive”.
“But in terms of actual value, we’re always going to add that.”
Information security director for AMER and EMEAA Managed Hotels at IHG Matt Gregory (pictured above, second from right) added security is “a constant evolution, as the attacker changes we change as well”.
“The tools are there to make things easier and more effective”, he emphasised, but echoed the fact there is a need for “human engagement” and oversight.
Scott explained from his perspective “we’re still not ready to trust AI to take over. It’s not human in the loop, it’s humans on the loop: continuous review with AI as a partner until you feel confident enough it can take [on] those low-level tasks with 100% certainty it’s doing the right thing, not hallucinating and going outside its” boundaries.
“I don’t think we’re there on that yet.”
Smith added “you know the culture of your organisation, you can read the room, you know the nuances, you can bring that to the table”.
“AI can organise the data, but especially [for] intelligence, there has to be a human there bringing that perspective,” she concluded.
On the attack.
Elsewhere on the keynote agenda, Stuart McKenzie, MD EMEA for Google Cloud subsidiary Mandiant Consulting, provided insight from its latest intelligence on threat trends facing the audience of security professionals.
Noting “attackers have adopted AI wholeheartedly”, he explained “we’re beginning to get to the speed and scale of some of the attackers that we have to start thinking about how we’re going to deploy AI to get into those battles”.
“We have to move as fast as the attackers, if not faster, or they’ll be through our environments so quickly”.
McKenzie added the abilities of criminals used to be five-to-seven-years behind activity from nations, but this gap is now nearer 18 months.
The ability for criminal threat actors to identify vulnerabilities is now greater than it’s ever been
Stuart McKenzie, MD EMEA
Mandiant Consulting
This “step-change” in speed and scale has seen hackers “exploiting vulnerabilities more often than not before anyone is aware of them”.
“The ability for criminal threat actors to identify vulnerabilities is now greater than it’s ever been with the help that LLMs provide them. It’s given them an uplift in sophistication”.
However, he noted it is not all “doom and gloom”, with new technologies and tactics available to aid defenders too.
The post Feature: How do security pros view the AI revolution? appeared first on Mobile World Live.
Login
