Artificial intelligence is changing the cybersecurity landscape, giving organisations new ways to identify threats while also helping cybercriminals automate parts of their operations. As attacks become faster and more sophisticated, security companies are increasingly adopting automated monitoring, deception technologies and AI-assisted vulnerability assessments to strengthen digital protection.
Industry experts warn that organisations have less time to identify weaknesses and respond to suspicious activity. The growing use of AI in cybercrime is also increasing pressure on businesses to strengthen their security systems before vulnerabilities are exploited.
How Are Cybersecurity Firms Using Decoy Systems Against AI Attackers?
One emerging defensive approach involves cybersecurity deception, in which organisations create artificial digital assets designed to attract suspicious activity and alert security teams.
Horizon3 Chief Executive Snehal Antani has highlighted the potential of deliberately planted decoy files and fake credentials to identify automated attacks. These assets are designed to help security teams detect unauthorised access attempts before attackers reach important business information.
Research published by Horizon3 in June reportedly tested 21 AI models, including models developed by OpenAI and Anthropic. According to the company's findings, the models were deceived by planted traps in 78.5% of tests, compared with 37% for human hackers.
Some advanced models reportedly fell for the decoys in more than 90% of tests.
However, deception technologies are not a complete security solution. Experienced attackers may recognise artificial assets, making it necessary for organisations to combine these systems with security updates, access restrictions and continuous monitoring.
How Quickly Can AI-Assisted Cyberattacks Develop?
The speed of automated cyber activity is becoming a major concern for businesses.
Research by cybersecurity firm Ensign InfoSecurity reportedly found that 10 advanced AI models successfully breached a test environment during initial assessments. However, their performance declined in later stages, particularly when attempting to obtain credentials and move between different systems.
Palo Alto Networks also disclosed an incident involving an attacker who reportedly used an advanced AI model to breach an enterprise network autonomously in less than 10 hours on September 2. The company estimated that a comparable operation could have taken a group of human attackers approximately two weeks.
Microsoft separately reported in October that AI could significantly reduce the time required for activities such as identifying credentials, discovering sensitive information and gaining access to computer systems.
Fortinet's Singapore and Brunei country head, Jess Ng, warned that organisations now have a narrower window to identify and address vulnerabilities before they are exploited.
These developments underline the importance of continuous security monitoring and timely software updates, particularly for organisations operating complex digital infrastructure.
What New Risks Are Emerging From AI-Assisted Cybercrime?
AI tools are lowering some of the technical barriers associated with cybercrime. Automated systems can help individuals with limited programming experience understand technical information, prepare deceptive communications and support different stages of cyber operations.
One example cited in the report is EvilTokens, a platform associated with phishing-related services. It reportedly affected 10,000 organisations across 120 countries within six months before Microsoft and its partners disrupted the operation in September.
Security researchers have also raised concerns about AI agents being able to identify relationships between multiple software weaknesses. When considered together, vulnerabilities that may initially appear less serious can create broader security risks.
Another concern emerged following the disclosure of critical PHP vulnerabilities on October 21. Cloudflare reportedly observed hacking attempts reaching millions after the disclosure. However, the involvement of AI in those attempts had not been confirmed.
This distinction is important: an increase in automated cyber activity does not necessarily establish that AI was responsible for every incident.
How Are Cybersecurity Companies Strengthening Their Defences?
Security firms are increasingly developing AI-powered tools to help organisations identify weaknesses and assess their exposure before an incident occurs.
Horizon3, for example, uses AI-assisted technology to conduct authorised security assessments of organisational networks, identify vulnerabilities and evaluate whether corrective measures have been effective.
The company announced in August that it had raised US$250 million, including investment from Singapore's Economic Development Board, to support its expansion and cybersecurity development.
Industry representatives are also emphasising the importance of proactive security planning. Association of Information Security Professionals president Emil Tan has called for organisations to anticipate potential threats instead of relying primarily on responses after a breach has occurred.
For businesses, this approach involves combining automated detection with regular system reviews, employee awareness, access management and clearly defined incident-response procedures.
Organisations can strengthen their internal risk-management framework through Shunyatax's internal audit services, which can support broader reviews of internal controls, compliance processes and operational risks.
Why Businesses Need to Prepare for Faster Cyber Threats
The growing use of AI in cybersecurity presents both opportunities and challenges. While defensive tools can help identify suspicious activity more quickly, criminals are also adopting automation to increase the speed and reach of their operations.
Businesses should not rely on a single technology to protect their systems. Regular vulnerability assessments, timely security updates, restricted access to sensitive information and employee training remain important components of a comprehensive cybersecurity strategy.
As AI-powered tools become more widely available, the ability to detect threats early and respond effectively will become increasingly important for organisations of every size.