Generative artificial intelligence, which includes the creation of deep fake videos, poses a significant threat to election security, according to a recent federal bulletin issued by intelligence agencies. The bulletin, sent to law enforcement partners nationwide, warns that both foreign and domestic actors could use generative AI to influence and disrupt the upcoming 2024 election cycle. This technology allows for the creation of realistic deep fakes that can be used to spread misinformation and sow discord, potentially impacting the integrity of the electoral process. The bulletin highlights the potential for generative AI tools to exacerbate existing tensions, disrupt election processes, and target election infrastructure.
Director of National Intelligence Avril Haines echoed these concerns during a recent Senate Intelligence Committee hearing, emphasizing the ability of generative AI to produce authentic and tailored messaging on a large scale. She pointed out that foreign influence actors could exploit this technology to create deep fakes that are difficult to attribute, posing a significant challenge to election security. Despite the growing threat posed by generative AI, Haines assured Congress that the U.S. is better prepared for election security than ever before.
The bulletin cited a specific example of generative AI being used to impersonate President Joe Biden in a fake robocall during the New Hampshire primary in January. The fake audio message encouraged recipients to withhold their vote until the November general election, demonstrating the potential impact of AI-generated content on voter behavior. The bulletin also highlighted a case in India where an AI video influenced voters to support a specific candidate, underscoring the global reach of this technology and its potential to manipulate election outcomes.
In addition to influencing voter behavior, generative AI could also be leveraged to target election infrastructure, according to the bulletin. The technology could be used by threat actors, including violent extremists, to identify vulnerabilities in election systems, aggregate a list of potential targets, and provide tactical guidance for attacks. The bulletin warns that violent extremists have experimented with using AI chatbots to supplement their tactical information, although there is no evidence yet of them using this technology for election-related attacks.
Overall, the bulletin underscores the urgent need for enhanced cybersecurity measures to protect against the growing threat of generative AI in the upcoming election cycle. As technology continues to evolve, it is crucial for law enforcement and intelligence agencies to stay ahead of the curve and develop strategies to counter the potential misuse of AI for electoral interference. By raising awareness about the risks posed by generative AI, the bulletin aims to equip law enforcement partners with the knowledge and tools needed to safeguard the integrity of the electoral process.
