To err is human: the role of AI in cybersecurity

Like many other nations, Australia has not been immune to a surge in data breaches and cybersecurity incidents, impacting both public and private sectors. Compounding financial and reputational losses, these incidents have exposed sensitive personal and financial information, leaving decision-makers to rethink their approach to cybersecurity measures.

Amid evolving cyberthreats, organisations are increasingly exposed to risks like phishing, social engineering, DDoS attacks and ransomware. It’s easy to assume that the greatest threats to our digital security are lurking in the dark corners of the internet, carried out by sophisticated hackers and malicious software. However, the truth is that most cyber incidents are not the result of complex code or impenetrable firewalls being breached; rather, they are caused by human error.

The proliferation of artificial intelligence (AI) in the workplace has also introduced new avenues for human error in safeguarding sensitive information, creating an even larger environment for cybersecurity.

However, AI is also providing a remedy to complement cybersecurity measures, particularly when it comes to data privacy. Despite this, more than half of Australians (57%) believe AI creates more problems than it solves. As data-fuelled technologies like AI and Machine Learning (ML) enter the mainstream, stewards of customer data now have dual responsibilities. They must build privacy and security into products and processes, while keeping that same data accessible and useful. So how can businesses take a more holistic approach when it comes to transforming security while also adapting new AI processes?

AI-driven proactive defence is better than a cure

AI already permeates every facet of our digital lives, eliminating tedious manual tasks from email filtering to identifying anomalous behaviours through advanced threat detection capabilities. Fuelled by data, it has enabled a shift from reactive to proactive responses to cybersecurity. For example, by applying AI and ML, organisations can analyse data to predict system and application outages or anomalies even before they happen. Security teams can then use this knowledge to create an algorithm — or a set of instructions — to remedy the problem or start an alert process.

When it comes to cybersecurity incidents, people are our Achilles heel and Australian business leaders cannot afford to overlook the profound impact of human error as the predominant catalyst behind major cybersecurity challenges. In security terms, human error is defined as unintended actions or lack of action by employees and users which cause or spread cybersecurity incidents. For example, a hacker pretending to be an employee may persuade a technical support member to reveal a password, or a seemingly friendly email may contain suspicious links. These incidents not only cause major headaches for under-resourced IT teams, but can lead to serious data breaches.

Cybercriminals perpetually seek vulnerabilities or loopholes within an organisation’s networks and applications. Frequently, the weakest link can be attributed to human errors, fragile authentication systems, or inadequately safeguarded assets. Of course, no organisation can protect itself completely from cyberthreats — companies must all acknowledge and assume that their systems will be breached in some way.

To transform security comprehensively and integrate AI processes seamlessly, businesses must recognise the fundamental role of human error in cybersecurity threats. This demands a holistic approach that combines cutting-edge AI technologies with robust cybersecurity training and awareness programs, ultimately ensuring a safer and more resilient digital future for individuals and organisations across Australia and beyond.

Addressing the ‘three Ms’ of cybercrime

The majority of cybercrimes stem from the ‘three Ms’: mistakes, misconfigurations and mismanagement. Often more destructive and challenging to control, human error can lead to unsuspecting employees falling victim to phishing attacks, resulting in data leaks. For instance, authentic-looking emails containing dangerous attachments or hyperlinks often do the trick.

Misconfigurations that exist in software subsystems or components are yet a‘nother human factor that exposes organisations to cyberthreats. For example, running outdated software or unnecessary features and services, and glitches or gaps in setting up cloud services or access controls.

Finally, the industry grapples with mismanagement of data and security as cloud usage surges. Cyber attacks stemming from security mismanagement often emerge when IT security teams exclusively depend on passive or reactive measures, like firewalls and anti-malware systems, to address cybersecurity concerns.

For most businesses today, the first step to navigate the complexities of managing cybersecurity in their business is to start with awareness and training. It is crucial for businesses to create a security culture that not only protects themselves, but also their customers’ data from both internal and external risks. Needless to say, security awareness training and simulated phishing security tests are crucial in helping enterprises today reinforce their ‘human firewall’ through regular employee education on risk awareness and vigilance.

There has been a significant shift in data privacy dynamics, with technology now having the dual role of using and protecting data. AI-based automation plays a pivotal role in enhancing privacy protections, alongside emerging technologies like homomorphic encryption, differential privacy and federated learning.

These technologies empower businesses to handle demanding tasks with unmatched speed and accuracy, placing less pressure on IT security teams. By embracing these innovations and adopting a comprehensive cybersecurity awareness strategy, Australian organisations are better placed to navigate an evolving digital landscape, ensuring data protection and sustaining trust among their customers.

Image credit: iStock.com/hamzaturkkol