Home / Technology / Human Error, Not AI, is Top Security Risk Today
Human Error, Not AI, is Top Security Risk Today
1 Oct
Summary
- Experts identify human error as the primary security threat, not AI.
- AI security risks are often amplified by basic human oversights.
- Organizations must fix fundamental security issues before fearing AI.
AI security experts convened at the HumanX AI conference, asserting that human fallibility, not artificial intelligence, poses the greatest current security threat. Panelists indicated that sensational headlines about agentic AI systems are premature, with the real danger stemming from humans failing to implement basic security measures.
Leading figures like Diana Kelley and Eric Doerr stressed that organizations often overlook essential security practices such as enabling multi-factor authentication. They noted that sophisticated users can also be overly permissive with AI capabilities. Madeline Lawrence added that many high-profile AI security incidents exploit simple, known vulnerabilities that have existed for years.
While AI accelerates the pace of cyber threats, old security strategies are proving inadequate. Concerns exist about AI agents being over-permissioned, potentially exposing sensitive data, as seen in research on GitHub's workflows. Despite these risks, the core issues often trace back to human oversight and the failure to establish proper controls around AI deployment.
The experts acknowledged AI's potential to enhance security by detecting vulnerabilities rapidly. However, they cautioned against expecting AI to eliminate software bugs entirely. The ongoing development of new software, coupled with persistent human errors, suggests that cybersecurity professionals will remain essential for the foreseeable future.
