Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

AI: Cybersecurity threat or opportunity?

Security professionals must weigh the benefits and risks of AI as it becomes more sophisticated and widespread.

[Images: Adobe Stock / everythingpossible]

Profile photo of Stu Sjouwerman

Artificial intelligence (AI) is having a profound impact that can’t be underestimated, prompting the question of whether it is beneficial or a curse to cybersecurity. To gain a better understanding of AI’s potential, it’s essential to delve deeper into the related opportunities and challenges that lie ahead.

AI SPELLS CYBERSECURITY OPPORTUNITY

Discovering and responding to threats in real time is one of cybersecurity’s biggest challenges. It’s humanly impossible to manually monitor and analyze every security alarm, attack vector, and attack surface.

On the other hand, AI can help analyze, co-relate, and interpret volumes of data from diverse sources such as network logs, security software and devices, applications, and endpoints. AI can identify patterns, detect anomalies, and report on potential vulnerabilities and threats. AI has incredible benefits for incident response; it can provide a detailed impact assessment with likely root causes, using natural language and summaries. AI can help address human error as well, flagging suspicious emails, reducing misconfigurations and inconsistencies, and improving cybersecurity practices.

AI SPELLS CYBER THREAT

For threat actors and cybercriminals, AI acts as a force multiplier, serving as a powerful tool for creating highly sophisticated phishing emails. AI can also be leveraged to fabricate entirely new identities or create deepfakes (synthetic images, audio, and video) for the purpose of manipulation, deceit, or disinformation. It can profile targets based on their social media posts, location, demographics, political associations, etc. It can reverse-engineer software code, hunt for vulnerabilities, or develop malware. Some researchers believe that since AI systems are not bound by human capabilities, they could execute cyberattacks using entirely new methods, or using tactics previously considered unfeasible.

MALICIOUS USE OF AI MAY OVERSHADOW ITS DEFENSIVE USE

Most security practitioners (77%) have serious concerns about the malicious use of AI. What’s more, 62% believe that offensive AI will at some point outpace its defensive use. For example, AI-drafted phishing emails have already become practically impossible to detect and can easily bypass traditional email filters. Phishing attacks are also increasing in volume thanks to generative AI. Researchers have successfully created polymorphic malware just using ChatGPT, and multiple reports of attacks show hackers successfully outsmarting and compromising AI tools.

HOW TO MITIGATE THE THREATS POSED BY AI

AI-powered threats are on the rise and the risks are real. Below are some recommendations organizations can employ to help tackle this threat.

1. Increase Security Awareness Training

Phishing and social engineering attacks are the biggest risks posed by AI. In fact, we’ve seen a 135% increase in “novel social engineering attacks” since the introduction of ChatGPT.

Organizations should prioritize increasing employee awareness of AI-related threats, fostering positive security habits, and aiding in the identification, reporting, and mitigation of attacks driven by AI. Policies and procedures must dissuade employees from disclosing confidential or proprietary information while using generative AI tools.

2. Conduct Regular Assessments And Simulations

Audit your IT infrastructure, code, applications, and devices to fix and patch any vulnerabilities in your security posture and procedures before a rogue AI can take advantage. Run phishing simulation tests regularly to determine if employees are ready to defend against AI-based attacks or if they need more personalized coaching.

3. Invest In AI-powered Security Controls

Cybersecurity interventions powered by AI and machine learning algorithms offer an opportunity to level the playing field with threat actors, helping organizations improve their prediction, detection, response, and recovery capabilities. With their higher degree of speed and accuracy, AI-based tools can take over routine tasks such as analyzing network traffic or monitoring user behavior, while security talent can focus efforts on higher-priority tasks. This can help reduce the stress and burnout of security teams while addressing the problem of attrition and talent shortages to a degree.

4. Develop Incident Response Plans Including Attacks On AI

When an incident strikes, emotions run high, panic takes over and it’s difficult to think clearly. Therefore, it’s always sensible to be well prepared rather than rush into an incident unprepared. It’s also possible that threat actors attempt to tamper with existing AI algorithms and datasets to manipulate their decisions or introduce flawed predictions.

When building or practicing incident response plans, be sure to cover scenarios where it becomes necessary to review, monitor, and recalibrate AI algorithms to avoid any biases, discriminatory outcomes, and privacy violations.

Bad actors will increasingly weaponize AI to boost their speed, scale, and accuracy of fraudulent and malicious activities. By harnessing AI-based security tools, prioritizing employee awareness of AI threats, and preparing the business for potential incidents, organizations will certainly be in a stronger position to confront the security challenges that lie ahead.


Stu Sjouwerman is the Founder and CEO of KnowBe4 Inc., the world’s largest Security Awareness Training and Simulated Phishing platform.