AI and cybersecurity: the 2025 implications

Concept artwork for AI and cybersecurity
generated by Dream by WOMBO.

Cybersecurity has always been concern for businesses, and any new technology could transform the battle. So far, AI is being used for both attack and defence, and can even be a target. 

Global tech executives are bracing for a significant rise in security threats alongside AI adoption, with 41% predicting a sharp increase in 2025, NetApp said.

"Data privacy and security concerns have remained top challenges globally year-over-year, with AI-leading countries like India, Japan, Singapore and A/NZ and the US ...facing nearly twice the number of security issues compared to AI-lagging countries," the company stated at the launch of its Data Complexity report.

Some 72% of Asia-Pacific (APAC) respondents said the rise in AI adoption had led to an increase in security challenges for their businesses, NetApp said. "AI-driven security challenges are weighing heavily on the minds of global tech leaders, with 59% identifying these threats as a leading global stressor.

"Likewise in APAC, over half of the surveyed respondents (54%) have recognised AI-driven security challenges as their top stressor. This heightened focus reflects the growing complexity of cyber risks."

Kevin Kirkwood, CISO, Exabeam, predicted that the battle between AI-weaponised attackers and AI-powered defenders will intensify. He said: "Malicious actors will increasingly use generative AI to create morphing malware—code that adapts and mutates to evade detection, making traditional defenses obsolete. These new strains of AI-generated malware will be more efficient and harder to trace. At the same time, defenders will lean on AI tools to streamline threat detection, asking more sophisticated questions and flagging abnormal behaviour more quickly."

"We’ll see an essential pivot towards AI security that cannot be ignored," agreed Reuben Koh, Director of Security Strategy, Akamai Technologies APJ. 

"Organisations will need to focus on two key areas: safeguarding their AI systems from vulnerabilities and defending against increasingly sophisticated AI-driven attacks. Cybercriminals are already leveraging AI to make their methods more effective and evasive, which means businesses must ramp up their security frameworks to counter these evolving threats." 

Koh from Akamai predicted that 2025 will see an increasing focus on protecting AI systems as well as defending against AI-driven attacks. "After all, cybercriminals are also looking at how to leverage AI to make their attacks more evasive, more efficient, and more effective. AI will lower the barriers to entry for attackers, accelerating their ability to identify and exploit vulnerabilities," he said.

AI-driven attacks

AI-driven phishing attacks will surge, said DigiCert, a global provider of digital trust. "The proliferation of AI will fuel an unprecedented surge in sophisticated phishing attacks, making them harder to detect. Attackers will leverage AI to craft highly personalised and convincing phishing campaigns, while automated tools will enable them to scale attacks at an alarming rate, targeting individuals and organisations with precision," said the company in a list of Asia-Pacific predictions for 2025. 

Pure Storage's 2025 predictions also touched on AI and phishing. "When gen AI was first introduced, we saw how ChatGPT was used to improve the quality of phishing emails. Cybercriminals have become even more sophisticated today, using recursive AI to find vulnerabilities in their target’s IT infrastructure," the company said. 

Prompt hacking

Steve Povolny, Senior Director, Security Research & Competitive Intelligence, Exabeam, agreed that AI will make it easier to launch cyberattacks. "You won’t need to be a coder to create sophisticated malware in 2025—AI will do it for you. Generative AI models trained specifically to generate malicious code will proliferate in underground markets, making it possible for anyone with access to deploy ransomware, spyware and other types of malware with little effort," he said. 

"These 'hacker-in-a-box' tools will automate everything from writing to deploying attacks, democratising cybercrime and increasing the volume and diversity of threats."

Nanhi Singh, Chief Customer Officer and GM, Application Security, Imperva, had a prediction in the same direction. "The capability presented by gen AI means that even inexperienced operators, who previously needed some degree of technical skill or knowledge, are increasingly able to launch sophisticated attacks with minimal effort," he warned. 

"Cybersecurity professionals as a result will face significant challenges in protecting their networks. Imagine a cyberattack tool that requires only the name of a corporate target to set off a series of malicious activities. Thanks to gen AI, it could well happen in 2025. Threat actors could use it to automatically generate and send phishing emails, and then once inside a target network, can leverage the technology to gain further levels of access. The ease of use and effectiveness of these tools will lead to an increase in the volume and sophistication of cyberattacks.” 

AI-driven defence

Steve Wilson, Chief Product Officer, Exabeam, said AI could become ubiquitous in security operations. "While AI has become common in tools like SIEM and UEBA, 2025 will see generative AI augmenting nearly every layer of cybersecurity, from endpoint protection to threat intelligence. The ubiquity of these systems will enable a much more dynamic and resilient security posture, able to address complex threat landscapes with speed and precision," he said. 

"Organisations should invest now in AI security certifications and frameworks, positioning themselves to meet emerging regulatory and compliance requirements around AI in cybersecurity. This foundation will be key as more AI systems are integrated across security architectures."  

"On the defensive front, AI copilots will become indispensable in cybersecurity operations, speeding up threat detection, investigation, and response. By 2025, every cybersecurity operator will likely be equipped with a generative AI copilot, streamlining complex analyses and providing actionable insights in real-time," Wilson added. 

"Companies should prepare to integrate these copilots, ensuring interoperability with existing security infrastructure and training operators to collaborate effectively with AI assistance. This dual human-AI approach will elevate response speed and precision, especially in high-stakes incidents."

Rahul Yadav, Chief Technical Officer, Milestone Systems, said that AI will enable security systems to go from analysis to action. "While traditional metrics like IQ measure cognitive ability and EQ gauges emotional awareness, a new capability is emerging: the power to act intelligently and autonomously — AQ (action quotient). Think of Tesla's self-driving cars, which don't just process road conditions, they smoothly navigate complex traffic scenarios in real time," he said.

"This shift toward action intelligence is particularly relevant in security operations. Traditional monitoring systems alert operators to potential issues, requiring human intervention for every response. In contrast, high-AQ systems can assess situations, initiate appropriate responses, and adjust their actions based on changing conditions. This capability will transform how we approach security management, making systems more proactive and less dependent on constant human oversight.

"The implications extend far beyond simple automation. These systems will be able to coordinate complex responses across multiple subsystems, from access control to emergency communications, creating more comprehensive and effective security solutions. The key is that these actions aren't just preprogrammed responses — they're intelligent decisions based on real-time analysis and learned patterns." 

Mastercard noted that companies are training AI models to predict and neutralise threats in real time. "Mastercard’s Decision Intelligence Pro uses gen AI to scan 1 trillion data points to predict in less than 50 milliseconds whether a transaction is likely to be genuine or not, boosting fraud protection rates by an average of 20% and as much as 300% in some instances," the company said.

Gen AI refers to generative AI, and SIEM is an acronym for security information and event management, while UEBA stands for user and entity behaviour analytics.

The role of the data centre

The increasing frequency and severity of ransomware attacks is driving a new, broader look at cybersecurity processes and the role the data centre community plays in preventing such attacks, said Vertiv in a list of 2025 predictions.  

"One-third of all attacks last year involved some form of ransomware or extortion, and today’s bad actors are leveraging AI tools to ramp up their assaults, cast a wider net, and deploy more sophisticated approaches. Attacks increasingly start with an AI-supported hack of control systems, embedded devices or connected hardware and infrastructure systems that are not always built to meet the same security requirements as other network components. Without proper diligence, even the most sophisticated data centre can be rendered useless," said Vertiv.

"As cybercriminals continue to leverage AI to increase the frequency of attacks, cybersecurity experts, network administrators and data centre operators will need to keep pace by developing their own sophisticated AI security technologies. While the fundamentals and best practices of defence-in-depth and extreme diligence remain the same, the shifting nature, source and frequency of attacks add nuance to modern cybersecurity efforts."

AI as a target

There will be an acceleration of attacks on AI, according to Jeffrey Kok, VP of Solution Engineers for Asia Pacific and Japan at CyberArk. "AI systems are increasingly attractive targets for cyber attackers due to the low barrier to entry and high likelihood of success. Many of the current AI models and implementations may not have been designed with adequate security protection and guardrails," he observed. 

"Organisations must prioritise stronger security measures and embed security frameworks directly into AI models to mitigate these risks."

Koh from Akamai said: "The hype surrounding large language models (LLMs) will face a stark reality check as security vulnerabilities come to the forefront. While some exploitable flaws have already been publicly disclosed, we can expect an uptick in both the frequency and severity of these issues. LLMs present a significant attack surface for malicious actors, and the nature and locations of these vulnerabilities will become increasingly evident. 

"As the risks become clearer, organisations will need to weigh the promise of LLMs against the potential security pitfalls, leading to a more cautious approach in their AI strategies."

The need for a holistic approach

"Organisations that fail to adapt to this AI-driven threat landscape risk severe financial losses, reputational damage, and potential business failure. Proactive investment in advanced cybersecurity measures and recovery strategies will be crucial for survival in the face of these evolving threats," forecast Pure Storage. 

Source: Accenture. Shankar.

Holistic cybersecurity strategies in 2025, said Vinod Shankar, Security Lead, Southeast Asia, Accenture. “With generative AI in the spotlight, cybersecurity is a parallel conversation that is taking centrestage. Accenture’s latest survey reveals that the top concerns of CXOs in APAC are related to various emerging cyberthreats including ransomware and deepfakes, with only 48% in APAC feeling completely prepared to defend themselves, as compared to 57% in Europe and 54% in North America.

"Lack of cybersecurity represents a serious business vulnerability, exacerbated by the rapid growth of AI implementation and deployment. Integrating and scaling AI exposes new attack surfaces in an already expanding threat landscape and increasingly complex technological landscape. Generative AI itself also allows malicious actors to develop and deploy more sophisticated and frequent attacks.

"To mitigate these risks, businesses must invest in advanced cybersecurity technologies, building comprehensive protection systems from the digital core to the supply chain. This includes leveraging the power of gen AI for defense, implementing quantum-proof security measures, and taking a human-centric approach towards sharing the responsibility of cybersecurity across the organisation.”

AI specialists needed 

As companies work to upgrade their security in 2025. traditional security operations centre (SOC) analysts could become obsolete, according to Exabeam. "The future of cybersecurity jobs will hinge on human expertise paired with AI innovation," said Gabrielle Hempel, Customer Solutions Engineer, Exabeam. 

"In 2025, the need for traditional SOC analyst roles will rapidly decline as AI and machine learning take over routine security tasks. Organisations will prioritise hiring AI specialists who can interpret, manage and guide advanced AI-driven security systems. 

"Threat-hunting roles will surge in demand, as human expertise is needed to contextualise and act on AI-generated insights. Companies will no longer rely on generalist cybersecurity teams but instead seek highly specialised professionals to stay ahead of increasingly-sophisticated AI-powered attacks."

Thales, on the other hand, said that AI tools have to be used right. “Cybersecurity vendors are increasingly integrating AI-assisted copilots to enhance their services for customers. These tools are great for helping to fill talent shortage gaps, which the ISC currently estimates at 4.8 million worldwide, but aren’t a replacement for internal teams. 

"In (2025), it will be less about the adoption of these tools and more about how security teams leverage AI tools' capabilities. Those looking to remain agile will likely utilize these tools to bring their threat investigation abilities to the next level,” said Todd Moore, VP, Data Security Products, Thales.

Recalibration

The hype will deflate, said BeyondTrust experts. "AI², or the 'artificial inflation' of artificial intelligence, is set to see its hype deflate across industries," the company said in a list of 2025 predictions. 

"While AI will remain useful for basic automation and workflows, much of the overpromised capabilities, particularly in security, will fall short in 2025. The focus will shift toward practical AI applications that enhance security without overwhelming organisations with marketing noise."

Koh emphasised that businesses should not forget security fundamentals. "Attackers will still come after our API endpoints, and they will still conduct phishing attacks. We need to ensure that patches are deployed in a timely manner, our safeguards are always on, and that people are continuously trained to identify and mitigate malicious activity. Because with or without AI, those threats are not going away."

Source: HPE. Harders.

Nick Harders, APJ Systems Engineering Director, HPE Aruba Networking, said that network security will also continue to be a priority, especially given the proliferation of AI-associated security risks. "The network will no longer just be a connectivity enabler, but also a security defender, serving as the foundation for built-in Zero Trust and SASE architectures to provide comprehensive network access security from the cloud to campuses and data centres."

SASE stands for secure access service edge.

"In five to 10 years, there will only be two types of companies: Those that leveraged AI to innovate, and those that no longer exist. With this harsh reality, CISOs must figure out how to be an enabler of AI, not a blocker. But with AI still in its infancy, very few have a strong understanding of the technology or the risks it may present… leading to extremely low levels of confidence that their organisation is well-prepared. The lack of understanding around AI is ultimately giving threat actors a leg up," concluded Grant Bourzikas, Cloudflare Chief Security Officer.

Explore

There are separate posts on AI in general, cybersecurity in general, as well as on agentic AI.

Hashtag: #2025Predictions