The Coming Code: How the Cyber Threat landscape will change in 2027
The cyber threat landscape in 2027 will continue to change along with the development of AI technologies - this is the preliminary forecast of experts. According to cybersecurity experts, attackers will increasingly automate their attacks using neural networks, search for new vulnerabilities, and use large language models to write various malicious software and conduct phishing campaigns. For more information about what cyber threats to expect in 2027, how dangerous they are and how to protect yourself from them, see the Izvestia article.
Which trends will become key in the field of cyber attacks
Various types of social engineering and phishing mailings will continue to be relevant for intruders attacking ordinary users in 2027, for example, to steal online banking login data, Dmitry Galov, head of Kaspersky GReAT in Russia and the CIS, told Izvestia during Kaspersky CyberSecurity Weekend 2026 in Tashkent. With the development of neural networks, they will spend less and less time creating convincing fakes and phishing emails.
"In terms of attacks on organizations, we should expect more changes, since AI technologies allow attackers to generate malicious code and complicate their schemes," says the specialist. — Neural networks conceptually do not change the techniques and tactics of intruders, but they scale and accelerate them. And this acceleration will create more and more cybersecurity problems for organizations.
According to Dmitry Galov, attacks on supply chains are likely to continue to gain momentum in 2027. Targets that are interesting to many groups of intruders are already well protected, so hackers have to get to them through contractors providing services or software. The expert calls the use of neural networks by organizations an additional risk: there are more and more complex and sometimes unpredictable AI systems within companies that increase the attack surfaces.
How hackers' tools will change in 2027
The number of successful cyber attacks in 2027 will grow by 30-40%, and well-known hacking methods are being further developed, according to Yana Avezova, a leading analyst at Positive Technologies. In particular, attackers will more actively use legitimate tools of administrators and information security specialists, which allow them to remain unnoticed for longer. The percentage of successful attacks using such software has already increased from 8% to 11%.
"In addition, attacks are expected to increase through internal update servers and centralized software management systems that allow malicious code to be distributed to multiple devices at once," says Izvestia's interlocutor. — Exploiting vulnerabilities in widespread Russian software will more often lead to simultaneous failures in the work of several organizations. In addition, attackers will more often impersonate representatives of departments, managers and technical services in order to obtain confidential data.
In turn, Sergey Tunchik, head of information security at the Digital Economy League, predicts that in 2027 one of the main cyber threats will be the use of autonomous AI agents for attacks. Such systems will be able to independently analyze available information, look for vulnerabilities in software, develop ways to exploit them and verify the result.
According to Sergey Tunchik, unlike traditional automation tools, AI agents adapt their actions in response to the results. For example, if the first method of penetration did not work, the agent can independently choose another option. In addition to autonomous AI agents, the expert also identifies adaptive malware among the potential cyber risks of the coming year, which changes behavior depending on the detected protection tools, which makes it difficult to identify and analyze the threat. Another likely target of hackers are AI agents of organizations: attackers can exploit vulnerabilities in systems with access to corporate data and services.
What will be the main targets of cyber attacks
The key targets for cybercriminals next year are likely to be the public sector, IT companies and industry, Dmitry Karasovsky, head of information security at Axiom JDK, told Izvestia. At the same time, IT companies and software developers will be attacked at a faster pace, since hacking one supplier opens the way to hundreds of its customers.
"Industry and critical information infrastructure facilities will continue to be subjected to attacks aimed at collecting information, disrupting operations and disabling systems," the expert predicts. — The financial sector will continue to face a high risk of fraud using voice and video tampering technologies.
Separately, the editorial interlocutor identifies two classes of goals that often remain outside the main risk assessments. Firstly, it is a small and medium-sized business through which large customers can be attacked. Secondly, the software development infrastructure includes code repositories and assembly systems, the compromise of which makes it possible to infect the finished product even before its release.
As for the peak periods for cyber attacks next year, the analytical data show that starting in 2022, the quarterly peak of confirmed incidents each time occurred in the second half of the year, says Director of Service Portfolio Development at Solar JSOC (Solar Group) Dmitry Ivanov. In addition, according to him, attackers traditionally use long weekends, New Year and May holidays when the response slows down. For trading and service companies, an additional window of risk is peak sales and the end of the year, when downtime is especially expensive.
— If we are talking about social engineering, then there is a link to some dates, but otherwise this rule no longer works, - adds Sergey Polunin, head of the IT infrastructure solutions protection group at Gazinformservice. — Using AI eliminates the need for planning and allows attackers to carry out continuous attacks all year round, adapting as they go.
How to effectively defend against cyber attacks in 2027
In 2027, organizations need to move from protecting individual systems to comprehensive monitoring of the entire IT infrastructure, according to Alexander Vinokurov, technical account manager at R-Vision. It is important not just to record individual security events, but to identify interrelated actions that may indicate the development of an attack.
— Artificial intelligence will play an increasingly important role in ensuring cybersecurity, — says Alexey Korobchenko, head of the information security department at the Security Code company, in an interview with Izvestia. — Information security products with self-learning modules will help to more effectively monitor anomalies in user behavior, traffic and other aspects of cybersecurity.
According to Dmitry Isaev, CEO of Stakhanovets, an expert in the field of IT and information security, the most mature approach in terms of information security today is to constantly check whether the behavior of an employee or an AI agent corresponds to his role, official duties and tasks. Local AI models analyze dozens of user activity parameters on workstations and form a behavioral profile. This is not surveillance, but the identification of deviations from normal behavior, the expert notes, since such deviations serve as an early indicator of risk long before a leak occurs.
— In 2027, humans and their digital counterpart, the AI agent, will remain both the most vulnerable and the strongest part of the defense. An employee's mistake can pave the way for an attack, but a competent specialist can stop it. The task of technology is to give this specialist the necessary data and support, and not to make decisions instead of him," concludes the editorial interlocutor.
Переведено сервисом «Яндекс Переводчик»