AI and cybersecurity are becoming closely connected

Artificial intelligence is changing how organisations create documents, analyse data, write software and support customers. It is also changing cybersecurity.

AI can help security teams identify suspicious behaviour, investigate incidents and discover software vulnerabilities. At the same time, criminals can use similar technology to create convincing phishing messages, automate reconnaissance and improve malicious software.

As a result, AI is becoming both a cybersecurity tool and a cybersecurity risk.

A New Level of Industry Coordination

One of the latest significant IT developments is the creation of a United States government coordination group connecting artificial intelligence developers with organisations responsible for essential services.

Announced in July 2026, the initiative is intended to improve the sharing of information about software and infrastructure vulnerabilities discovered by advanced AI systems. Organisations expected to participate include major AI and technology companies, while government involvement spans national security, defence, finance and cybersecurity agencies.

Although the initiative is based in the United States, its importance is international. The software, cloud platforms and AI models involved are used by organisations throughout the world, including Australian businesses.

The announcement also reflects a broader change in attitude. AI security is no longer being treated solely as a matter for individual software vendors. Governments and major technology providers increasingly view it as an infrastructure issue requiring coordinated disclosure and response.

How AI Can Help Defenders

Modern IT environments generate enormous volumes of information. Security teams may need to examine authentication attempts, firewall records, network traffic, application activity and alerts from thousands of devices.

AI systems can assist by identifying unusual patterns and prioritising the incidents most likely to require human attention.

Potential uses include:

• Detecting suspicious login behaviour
• Identifying unusual network traffic
• Classifying phishing messages
• Analysing potentially malicious files
• Summarising security alerts
• Searching software for vulnerabilities
• Helping incident response teams understand an attack
• Automating repetitive investigation tasks

Recent research into cybersecurity powered by AI highlights areas including intrusion detection, malware analysis, phishing detection, natural language processing and collaborative security methods that preserve privacy.

The principal benefit is speed. AI can review more information than a human analyst could examine manually and may help security staff focus on the most significant threats.

How Criminals Can Use the Same Technology

The same capabilities can be misused.

Generative AI can produce polished messages with correct spelling, appropriate formatting and convincing business language. This makes it easier to create phishing emails that appear to come from a manager, supplier, bank or service provider.

AI can also help attackers:

• Research potential targets
• Translate scams into multiple languages
• Create personalised fraudulent messages
• Modify malicious code
• Search for exposed systems
• Generate fake voices, images or video
• Automate parts of an attack

This does not mean every cyberattack is now completely controlled by AI. Human attackers still need objectives, access and technical knowledge. However, AI can reduce the time and effort required to perform certain tasks.

Why Small Businesses Should Pay Attention

Large organisations usually have dedicated security teams, monitoring platforms and formal incident response procedures. Small businesses may depend on one internal IT employee or an external service provider.

This can make smaller organisations attractive targets, particularly when they rely on outdated systems, shared passwords or informal approval processes.

The rise of attacks enhanced by AI makes basic security practices even more important. Businesses should:

1. Enable multifactor authentication on important accounts
2. Maintain reliable and tested backups
3. Install operating system and application updates promptly
4. Remove accounts belonging to former employees
5. Use separate administrator and everyday accounts
6. Train employees to verify unusual requests
7. Require a second form of confirmation for payment or banking changes
8. Maintain an incident response contact list

A convincing email or voice message should not be enough to authorise a sensitive transaction. Requests involving money, passwords or confidential information should be confirmed through a separate, trusted communication method.

AI Tools Also Need Security Controls

Businesses adopting AI services should assess them in the same way they assess other cloud applications.

Before employees upload information to an AI system, the organisation should understand:

• Whether submitted data is retained
• Whether it may be used to improve the service
• Where the information is processed
• Who can access the organisation’s account
• Whether activity can be logged and audited
• What information employees are prohibited from entering

Confidential client records, passwords, internal financial information and proprietary source code should not be pasted into public AI tools without appropriate approval and protection.

Organisations also need to manage output generated by AI carefully. AI systems can make mistakes, produce insecure code or present incorrect information confidently. Human review remains essential.

The Human Element Remains Critical

AI may improve threat detection, but it does not replace good security management.

A company can invest in advanced security software and still be compromised because an employee reused a password, approved a fraudulent payment or ignored a software update. Technology is only one layer of protection.

The strongest approach combines:

• Secure systems
• Clear policies
• Regular employee training
• Effective access controls
• Tested backups
• Human verification
• Continuous monitoring

What Comes Next?

AI developers, governments and cybersecurity providers are likely to share more information about vulnerabilities discovered through advanced models. Organisations may also begin using AI more extensively to examine software before it is released.

At the same time, attackers will continue experimenting with automation, impersonation and increasingly personalised scams.

The objective should not be to avoid AI entirely. It should be to adopt it deliberately, understand how information is handled and maintain security controls that do not depend on users recognising every sophisticated attack.

AI is rapidly becoming part of everyday IT. Cybersecurity must become part of every AI decision as well.