Concerns Rise Over AI Disinformation as U.S. Presidential Election Approaches

With the upcoming presidential election in November, worries regarding AI-generated disinformation are on the rise in the U.S.

A Pew Research Center survey conducted on September 19 indicates that almost 60% of Americans, with a nearly equal divide between Democrats and Republicans, express “extreme” or “very” high concern about the potential for AI to produce false narratives about candidates. Additionally, around 40% of respondents believe that AI will be predominantly misused during the election process.

Following this, a U.S. intelligence official reported on September 24 that both Russia and Iran have been utilizing AI to manipulate videos of Vice President Kamala Harris as part of their election influence tactics.

Nvidia’s CEO, Jensen Huang, urged the U.S. government to not only regulate AI but also take the lead in its advancement. He proposed that government agencies like the Department of Energy and the Department of Defense actively leverage AI technologies, suggesting the establishment of an “AI supercomputer” to enhance research and innovation.

Huang also warned that the sector’s energy requirements are set to increase significantly due to the larger amounts of data needed for training AI models. He emphasized that future AI systems will depend on existing models for their training, necessitating data centers to consume 10 to 20 times more energy than they currently do. To address this surge in energy demand, he recommended constructing data centers near abundant energy sources, as the location of learning is not a concern for AI systems.