Protecting Ourselves from AI Cyber Hacking
Student Summary by Emma Seo
Original Source: Kathryn Hulick, Science News, June 26, 2026
Image created by Emma Seo
Abstract
Background: As AI is growing in advancement, it is becoming better at hacking. People can use
AI to steal someone's identity, disrupt core operations of banking and health care systems, and
much more. If cyber security defenders aren't prepared, problems might be caused by cyber
attackers with the help of AI.
Objective: This article draws on warnings from Five Eyes and on an interview with AI security
expert Michael Alexander Riegler. Five Eyes is a secretive intelligence alliance of five countries
(Australia, Canada, New Zealand, the UK, and the US) that works to gather intelligence and
respond to security threats.
Method: Science News interviewed Riegler, an AI security expert, to gather information about
risks of AI and the reality behind it.
Results: The information shared from the interview provides deeper insight into the dangers of
AI. Not only is the AI model causing the security risk, but also the systems around it. Other
causes are whether the model can access the internet, whether it can test its own code, or what
tools we provide it. To maximize protection from complex cyber attacks, we can have our
software up to date, use two-factor authentication, etc. There is also an upside to having access to
the same tools as attackers. We can use AI to find security holes and close it to prevent attacks.
The interview also covered how companies are dealing with AI security risks.
Conclusion: Through this interview, people gained deeper insight into the dangers, protection
methods, and advantages from attacks with AI involved.
