President Joe Biden was apparently impressed, and a little concerned, by the capabilities of generative AI.
Chip Somodevilla/Getty Images
After over three decades in the Senate, eight years as vice president, and three presidential campaigns, you'd think nothing would surprise President Joe Biden.
Then, last spring, he tried out ChatGPT. A few months later, he signed sweeping legislation targeting the new technology.
Arati Prabhakar, Biden's chief science and technology advisor and director of the White House Office of Science and Technology Policy, told Wired that she and Biden put the bot up to a few tasks.
First, they asked it to explain a lawsuit between Delaware (the state Biden represented as a senator) and New Jersey (the home state of singer-songwriter Bruce Springsteen, to whom Biden had just presented the National Medal of Arts) as if it was talking to a first grader. "OK, kiddo," the bot began.
Bloomberg/Getty Images
Then, they asked it to write a legal brief for a Supreme Court case, write a song in the style of Springsteen, and generate an image of Biden's dog, Commander, in the Oval Office.
"Wow, I can't believe it could do that," Biden told Prabhakar, according to Wired.
But his first encounter with generative AI also sparked concern.
Prabhakar told Wired that Biden later asked the team to address the potential risks of AI. That led to the sweeping executive order he signed in October. The order requires major tech companies to adhere to certain safety guidelines, notify the federal government of their work, and share testing results.
Biden also reportedly told his cabinet that AI would touch the work of every department and agency at a meeting in early October. "The rest of the world is looking to us to lead the way," he said.
