President Joe Biden has been cautiously optimistic about artificial intelligence.
“AI promises an enormous, enormous promise of risk to our society and our economy and our national security,” the president said over the summer. “But also incredible opportunities.”
The president’s cautious optimism morphed into a new executive order Monday.
“Safe, secure and trustworthy development of the use of artificial intelligence,” he said as he signed the order which requires new safety assessments, guidelines in the areas of civil rights and equity and further research on how the labor market may be impacted by AI.
“It’s about time,” said Rensselaer Polytechnic Institute professor Jim Hendler.
The data scientist leads RPI’s Future of Computing Institute and is one of a number of experts in the field who believe regulating the technology sooner rather than later is critical.
“There has been controversy over whether it should be a new organization or strengthen current laws,” he said.
He feels the latter is the right approach and believes the executive order is a good start.
The White House reports a number of companies leading the AI charge are committed to the executive order, which requires them to share safety test results with the federal government and will create guidance for watermarking to help the public better identify AI-generated material.
“That really is important because it’s not saying you can’t do this or you should do this or you must do this but saying let's have a policy in place how to do risk assessment,” Hendler explained.
The professor wishes the executive order addressed building a better workforce on the government level.
“Finding a level of employment that makes sense for people who understand this technology,” he said. “Who can really help the government deal with it is going to be very important.”
People like Inwon Kang, a doctoral student at RPI, is focusing his coursework on AI.
“It looks promising,” he said. “As long as we can act on it.”
Kang said the issue of privacy is one that comes up often which is included in the executive order.
“It’s great that they cover the privacy part in terms of protecting people’s data and making sure we know what data the model was trained on,” he said.
Kang calls this an encouraging moment and hopes it results in more people familiarizing themselves with AI and its potential.
“Not everyone needs to be an expert,” he said. “It’s like a car. I couldn’t tell you how a car runs itself but I know how to drive it in such a way that I don’t cause any harm to people or myself.”
On the heels of this executive order, leaders from around the world will gather to discuss this very topic, a summit of sorts organized by the British prime minister. The European Union has actually been more proactive in staying ahead of AI and its potential negative implications as it prepares to pass sweeping legislation while congressional leaders in Washington are just getting started.