The Department of Energy (DoE) is pursuing tech policy and workforce training initiative aimed at illuminating ethical concerns with the use of artificial intelligence (AI) technologies and making sure that DoE’s AI projects work to maintain human dignity and do not cause harm, an agency official said.
Pamela Isom, director of the AI Technology Office (AITO) at the DoE, explained that AI doesn’t have awareness of itself. It can only separate right from wrong based on data that has the label “right” and the label “wrong” attached to it, she said.
“The only moral compass there is when talking about AI, is that of its developer who then sets the bar for what is right and what is wrong,” Isom said during a virtual event hosted by Cognilytica on Feb. 17. “Ethical AI is designed and deployed to deliver equality, fairness, justice, safety, and integrity.”
“Ethical AI can save lives, but unethical AI can lead to an erosion in public trust and the slow progress and adoption of AI,” Isom continued.
To ensure the utilization of ethical AI the agency has deployed several practices surrounding two crucial areas: test and evaluation, and workforce training and diversity.
According to Isom, DoE is ensuring its workforce is educated and certified on AI ethics. To do that, the agency has facilitated and supported workforce training on best practices for developing and utilizing ethical AI – including establishing protocols that can help find and mitigate bias throughout the AI lifecycle to ensure equitable outputs.
DoE also has established protocols to help find and mitigate bias throughout the AI lifecycle, including AI independent verification and validation teams & tools.
“We can proactively detect, assess, and mitigate bias in AI systems,” Isom said.
In addition, DoE is developing and establishing tools to audit and certify AI systems as responsible and trustworthy.
“AI is here to stay. It increasingly informs high impact decisions, jobs, healthcare, banking, energy distribution, the justice system, etc.,” Isom said. “So, we are accelerating the research, development, deployment, demonstration, and adoption of responsible and trustworthy AI.”