Telcos depend on AI to thwart hackers

by Ken Wieland

23 July 2019

LONDON — Of all the industry sectors covered by a recent cybersecurity report from the Capgemini Research Institute, telecom looks the most enthusiastic in embracing AI to fend off malicious attacks. Four out of every five IT security specialists working in this sector, according to survey conducted by Capgemini, said they counted on clever AI algorithms to help identify threats and thwart attacks.

Other sectors covered by the report — Reinventing Cybersecurity with Artificial Intelligence: the new frontier in digital security — were automotive, banking, consumer products, insurance, retail and utilities. On average, less than 70% of all the 850 senior IT security specialists canvassed — across the seven different sectors — thought they would not able to respond to cyberattacks without AI. 

One reason why telcos rely so heavily on AI is they have much more to lose than most in the event of a data breach. This focuses minds. When asked about the approximate financial damage to their organisation resulting from a cybersecurity breach, 40% of IT security specialists working in the telecom sector put it at over $50 million. This was a much higher proportion than any other industry sector covered in the report. (Consumer products was the nearest with 26% of respondents saying that financial damage topped $50m.) By having huge amounts of customer data, said Capgemini, telecom firms were an ideal target for cyberattacks

Related: A Telco Executive’s Guide to AI

Putting a strain on telcos is exponential increases on traffic across their networks, which makes it extra difficult for cyber analysts to spot changing patterns of behaviour. It’s why US heavyweight AT&T is using machine learning to detect new patterns in network traffic and help root out ‘bad actors’ that can cause network disruptions or data breaches.

“I see more than 100 billion potential vulnerability scans and probes across our global backbone every single day,” said Bill O’Hern, senior vice president and chief security officer at AT&T. Referring to device connectivity, software-defined networks and 5G, he added: “You can’t really protect these new technologies with a legacy approach. It’s really key to understand that you need to evolve your capability along with these technologies and take advantage of machine learning or other AI technologies in harmony as a platform across your whole ecosystem.” To compile risk-rating scores in the network, Capgemini recommends that telcos use metrics that are data-driven and quantitative, rather than depending on domain insights from cyber analysts.