Artificial intelligence (AI) has emerged as both a significant threat and a crucial tool in the fight against cybercrime. While AI has greatly augmented threat
actors’ attacks, a new study by IBM said AI tools can shorten the lifecycle of a data breach and reduce costs for businesses by as much as $2.84 million.
In its annual Cost of a Breach Report, IBM found that Canadian organizations paid, on average, about $6.32 million per data breach, down from $6.93 million in 2023.
Financial services and technology companies paid the most, with breaches averaging $9.28 million and $7.84 million, respectively. Phishing attacks and stolen or compromised credentials were the most common initial attack vectors, accounting for 16% and 15% of all breaches, respectively.
Encouragingly, two in three (61%) companies globally have invested in security AI and automation to combat breaches. IBM said organizations equipped with these technologies had breach lifecycles that were 54 days shorter.
“Years ago, we predicted AI would be impactful, and this report confirms that,” said Daina Proctor (pictured), cyber security service line delivery leader at IBM Canada. “It absolutely shows that you’ve got a shorter breach lifecycle, and it brings down the cost.”
With an average breach cost of $6.3 million, investing in AI and automation to reduce that cost by nearly half “seems highly worthwhile,” said Proctor.
Leveraging AI, machine learning, and automation into cybersecurity strategies is increasingly needed. But it’s not enough to simply invest in AI; companies must understand how to deploy it effectively to get the best return on investment.
“From my experience, it’s important to walk before you run,” said Proctor. “Start with machine learning to automate repetitive tasks, then move to broader automation where you’re confident in the outcomes.”
However, the deployment of AI is not without risk. Organizations need to ensure that the data powering their AI systems is secure and that the AI models themselves are free from bias.
Generative AI (gen AI) is increasingly popular and can have a broad range of use cases. Proctor called on organizations to secure large language models, check for bias, and tailor applications to the right audience.
“The proliferation of AI across security programs and organizations is clearly beneficial,” she said. “However, it’s crucial to ensure the data we’re using within AI is properly secured to avoid creating new risks as we increasingly rely on these technologies.”
On the other hand, cyber threats continue to evolve, with malicious actors leveraging the same AI capabilities that defenders use. This creates a constant arms race between cybersecurity professionals and cybercriminals.
According to the IBM report, the most expensive attack vectors were business email compromise, phishing, social engineering and stolen or compromised credentials. It noted that gen AI may be playing a role in creating some of these phishing attacks, allowing even non-English speakers to produce grammatically correct and plausible phishing messages.
“The report and conversations with organizations confirm that AI is driving much of the current threat activity,” Proctor said. “This makes it even more critical for us to use AI to operate at the same speed and velocity to counter these threats.”
The rising cost of data breaches has significant implications for the insurance industry. Unlike other sectors where risks can be quantified with a high degree of accuracy, cyber insurance remains an unpredictable and constantly shifting battleground. Proctor called it a “wild, wild west.”
“If you put your car in a garage, drive less than 10,000 kilometres a year, and know it’s a certain car make and model, you know that this is the risk that goes with you. That model for the cyber industry is just not there yet,” she told Insurance Business.
This unpredictability is a significant hurdle for carriers who must figure out how to protect their clients from ever-more sophisticated threats while also managing their exposure.
When it comes to risk management, Proctor lamented that organizations are still grappling with the same basic issues they have for years, though the sophistication of attacks has increased.
Despite the technological advancements and substantial investments being made in cybersecurity, the costs associated with data breaches globally are still climbing. IBM reported the global average cost of a data breach in 2024 reached $4.88 million, increasing 10% from the prior year.
And while average costs have lowered in Canada, the number of incidents has increased to about 27,000 breaches a year, or almost 75 breaches per day, an all-time high.
“The fixes are known. I’ve been in this industry for a little over 15 years now, and we’re still doing the same things,” said Proctor. “So, why aren’t we better off?”
What are your thoughts on AI and cybersecurity? Please share your comments below.
