CISA Chief Talks Artificial Intelligence Risks
Cybersecurity and Infrastructure Security Agency (CISA) Director Jen Easterly recently discussed the potential risks of artificial intelligence (AI) during the Modern Conflict and Emerging Threats Summit. She expressed concerns about how adversaries could use AI technology, such as OpenAI's ChatGPT, to carry out cyberattacks or deploy deep fakes. ChatGPT-5, a more powerful version of the technology, is expected to be released before the United States' 2024 election.
CISA plans to use AI for code translation and review. However, the agency has yet to come up with ways to eliminate the threats posed by language models. In response to these concerns, CEOs of leading AI companies met with President Joe Biden and other officials to commit to opening their models during the Defcon hacking conference in August and cooperating with the government's effort to address AI security risks.
The Department of Homeland Security has also established an internal AI Task Force to maximize AI's use while countering its risks. As AI's development continues to accelerate, security experts are urging caution and emphasizing the importance of addressing potential risks.
