The state should prioritize training for state government personnel as it plans how to deploy artificial intelligence in government operations, according to witnesses during a recent public hearing held by the Little Hoover Commission.
“We are currently in a position where we do not fully understand what all of the socio-technical risks of AI are moving forward,” said Brandie Nonnecke, founding director of the CITRIS Policy Lab at the University of California, Berkeley. “It’s going to be evolving as the technology changes. And we need to continue to equip… state workers with that training.”
The July 25 meeting was the Commission’s third hearing in its study on artificial intelligence. This hearing explored best practices for state government procurement of AI applications and mitigating risks.
Tom Andriola, Vice Chancellor for Information Technology and Data at UC-Irvine, stressed the importance of involving people from all industries and technological skill levels in conversations around the opportunities or potential risks of AI.
“Everyone needs to have these experiences because they’re going to navigate their lives going forward with these tools…that may be able to… advance their careers or change the way that they interact with family members,” said Andriola during his testimony. “We want everyone to have their own working understanding. As I like to say to people, AI is not taking your job, but the person who knows how to use these tools more effectively may be the one to take your job.”
Lila Tretikov, Partner and Head of AI Strategy, NEA, suggested using AI to help augment the state’s customer service functions.
“AI is very well-suited for customer support scenarios, requests, or feedback,” said Tretikov. “The fundamentals of artificial intelligence – which is translation, transcription, voice synthesis – are particularly useful in broad scenarios where a wide range of constituents are present, making government resources and government services much more inclusive and much more accessible to all citizens.”
California State Senator Ben Allen served as the closing witness and provided insight on current efforts to increase enrollment in government assistance programs using AI.
“Our own state low-income oversight board put out a report last year recommending a series of changes to the state’s regulated energy, water, and telecoms programs to increase enrollment,” said Senator Allen. “Among the proposals is the idea that we should increase targeted marketing and outreach to potentially eligible customers, that we should develop a concurrent enrollment program, …and that we really ought to capitalize on technology to streamline the process.”
The state also has the opportunity to influence the AI industry as it explores how to purchase and deploy the technology.
“Essentially, the way that the [state] departments and agencies are interpreting responsible AI governance is going to shape how industry builds and deploys their AI tools. The State of California has a large purchasing power. Vendors need to sell their products,” said Nonnecke.
The Commission’s first hearing on artificial intelligence, held in May, presented an overview of California’s efforts thus far to incorporate AI into State Government operations. The second hearing, held last month, assessed how artificial intelligence can help California state government better serve its vulnerable populations with a focus on addressing food insecurity.
The Commission expects to release reports later this year on the implementation of artificial intelligence in state operations and how artificial intelligence can be used to assist California’s vulnerable populations. To be notified electronically of meetings, events, or when these reviews are complete, please subscribe to Little Hoover Commission updates: https://bit.ly/LittleHooverNews.
AI: How the State Should Acquire Applications, and How to Assess the Risks
