Cybersecurity & AI

AI or Artificial Intelligence is a web-based tool used to complete processes or tasks that would typically take human intelligence to do. These tasks could include analyzing numbers, identifying objects, making decisions, and learning. Only in the last few years has AI gone truly mainstream, incorporating itself into people's daily lives. With it becoming so prevalent in our world today, we need to make sure we understand the risks it poses and how to guard against them from a security standpoint.

Where is AI?

Before examining the risks, we should first acknowledge that AI is everywhere today, taking many different forms, such as:

Social media

Smart Devices

Phones

Language Models

Risks with AI

While AI has been helpful in a lot of ways, there are also lots of risks when it comes to AI.

Misinformation

With AI photos and videos getting more advanced every day, a lot of what we see may not be as real as we think.

Privacy

When using AI, data is often stored, which puts personal information at risk of being breached or misused.

Overreliance

Many users use AI too often and have become overly dependent on it, and this can make it harder for them to learn new things as AI can do it all for them effortlessly.

Malicious attacks become easier

Attacks such as phishing and malware become easier to create. While most AI have safeguards around using them for malicious intent, these are often easy to get around.

Real life examples of AI Hacks

Real-world examples illustrate how easily AI systems can be manipulated or generate inaccurate claims. In December 2023, a Chevrolet dealership’s AI chatbot was reportedly tricked into offering a $76,000 Tahoe for just $1 after a user deliberately manipulated its responses. This incident highlights how customer-facing AI tools, commonly integrated into websites, can be vulnerable to relatively simple prompt-based exploits (Prompt Security, 2024). For a detailed account of this event, see the original post: https://x.com/ChrisJBakke/status/1736533308849443121.

In another case from April 2024, an AI chatbot falsely accused NBA player Klay Thompson of vandalizing property in Sacramento, California. Analysts suggested the system may have misinterpreted basketball-related language—such as “throwing bricks,” a term used to describe missed shots—resulting in a fabricated allegation (Olavsrud, 2025).

These examples underscore the risks associated with AI-generated content, particularly in situations where outputs may be misleading, manipulated, or taken out of context.

Being safer with AI

A few ways we can help to guard against the dangers AI poses are by going about adequate security measures ourselves. We can define boundaries so AI doesn't overstep, employ better authentication protection to verify users, and conduct regular security testing. This way we'll be staying ahead of the curb on the latest developments in technology and we'll be prepared for what comes next.