In this week’s bulletin, Charlie investigates the newest AI-driven scams and examines the advantages and disadvantages of AI in the online world with the help of Google Gemini.
This week, I was the allocated tutor for the BCT Certificate in Cyber Incident Management Course. This is the first time in a year and a half, so I was quite excited about acquainting myself with the course again and updating it with the latest cyber knowledge. The course runs around every three months, I always feel there is a need to update the course with the latest happenings in the cyber world. I decided to do some research to find out the latest thoughts on AI and cyber, as one of the challenges for the course participants was to include a slide on this topic.
Discussion in the course about AI was brought on by the story I noted in the news about a woman in Hong Kong who was reportedly the victim of a deep fake scam where she believed she was communicating with a high-ranking executive at her company. The scammers used deep fake technology to create videos and audio recordings that made it appear as if the executive was requesting urgent financial transfers. The woman, believing the deep fakes, transferred a significant amount of money to the scammers before realising she had been deceived. This is not the first time this scam has been attempted, but in the past, it’s known to have been a senior manager who have been asked to transfer money via email. Now, with AI, deep fakes’ voices and digital imagery can be used to attempt the scam.
I have wrote on this subject before in 2019; click here to read.
For this bulletin, I have decided to put AI against itself (Google Gemini). Here, I will explore both the positive and negative applications of AI in the cyberspace.
The positive uses of AI are described as follows:
- Enhanced Threat Detection: By analysing vast amounts of data in real-time, AI can identify anomalies and suspicious patterns that might escape human analysts. This allows for early detection of malware, phishing attempts, and other cyberattacks before they can cause significant damage.
- Automated Response: AI-powered systems can respond to threats instantaneously, patching vulnerabilities, blocking malicious traffic, and quarantining infected systems. This swiftness is critical in mitigating the impact of attacks, especially zero-day vulnerabilities where traditional patching approaches fall short.
- Predictive Security: AI algorithms can learn from past attack patterns and predict future threats. This proactive approach allows security teams to focus resources on areas most likely to be targeted, optimising their defensive posture.
- Personalised Security: AI can tailor security measures to individual users and their specific risk profiles. This allows for a more nuanced approach, avoiding unnecessary restrictions while safeguarding against individual vulnerabilities.
The same capabilities that make AI a powerful defensive tool can also be weaponized by attackers. The disadvantages and the offensive use of cyber are:
- Advanced Phishing: AI can be used to personalise phishing emails with convincing language and mimic user behaviour, making them harder to detect. This can trick even seasoned users into giving away sensitive information.
- Automated Attacks: AI can automate attack processes, launching large-scale campaigns with minimal human intervention. This can overwhelm defences and create widespread damage in a short time frame.
- Deep fakes and Social Engineering: AI can be used to create realistic deep fakes, manipulating video and audio to impersonate individuals and spread misinformation. Hence, the Hong Kong scam worked, and many of you would have seen the deep fake of President Biden telling voters not to vote in the forthcoming presidential election. This can be used for social engineering attacks, sowing discord, and compromising trust.
- Evolving Malware: AI-powered malware can dynamically adapt to counter security measures, making it more difficult to detect and eliminate. This poses a significant challenge for traditional signature-based defences.
This is an evolving landscape, and AI will be utilised on both sides to increase the cyber arms race.
