The AI Threat: Why We Haven’t Seen the Worst Yet, But It’s Coming

AI technology is revolutionizing industries, but it’s also empowering cybercriminals. From AI-generated deepfakes to elaborate scams, cyberattacks are evolving. While AI-based threats have yet to reach their peak, experts warn that the worst is just around the corner.

AI in Cybercrime Today

Cybercriminals already use AI for scams, such as fraudulent ads featuring deepfakes of celebrities or executives. Victims are lured into financial traps, particularly with cryptocurrency or stock manipulation schemes. This type of content manipulation — especially with AI-altered videos — is becoming increasingly common in the dark web.

Michael Price, CTO of digital protection firm ZeroFox, notes that malicious actors have begun advertising services to create videos featuring U.S. military personnel saying fabricated statements. This could fuel disinformation campaigns aimed at destabilizing geopolitical situations and undermining trust in institutions. These AI-powered videos are designed to influence public opinion rather than steal money directly.

Real-Time Deepfakes: The Looming Threat

The current state of AI threats, while significant, is still in its early stages. However, Price emphasizes that we are approaching a turning point, much like the rise of ransomware. Just as ransomware went from a niche threat to a global epidemic, real-time deepfakes could soon become a regular cyberattack method.

Real-time deepfakes allow cybercriminals to impersonate individuals on live video or phone calls. Imagine receiving a call from your CEO, who looks and sounds like the real person, but is actually an AI-generated impersonation. This could lead to unauthorized financial transfers or confidential information leaks. The technology hasn’t yet become widespread, but as it improves, the potential for damage grows.

Using AI to Defend Against AI

Fortunately, AI can also be used to defend against these evolving threats. Large language models (LLMs), speech-to-text software, and AI-enhanced translation tools are helping security teams analyze vast amounts of media and detect suspicious activity. AI allows defenders to interpret underground forums, media content, and other sources faster than ever, offering real-time insights into ongoing threats.

Price explains that AI tools will soon make it easier to monitor and assess digital content, from videos to radio transmissions, providing critical information that could help prevent attacks before they escalate.

Conclusion: Preparing for the Breaking Point

While we haven’t yet reached the breaking point of AI-based cyber threats, the rapid advancements in AI technology suggest that it’s only a matter of time. Businesses and individuals must stay vigilant, improving their cybersecurity awareness and adopting AI tools for protection. The key lies in recognizing the potential dangers early and adapting defenses before real-time deepfakes and other AI-powered attacks become a widespread reality.