During the morning keynote speech, many of the 938 reported attendees at the conference heard the latest about how AI will help their jobs, and potentially make them harder.
State tech leaders should never lose sight of the human factor when it comes to generative and other forms of artificial intelligence, said Theresa Payton, White House CIO under the George W. Bush administration.
During her presentation, she urged listeners to use previous lessons about putting the users at the center of plans as agencies now map out the AI future. Payton’s remarks were a reflection of the broader push for more human-centric design in government technology.
She cautioned that artificial intelligence technology, while helping public agencies with a variety of tasks both mundane and novel, also will provide much better access to cyber criminals.
For instance, ChatGPT can provide what amounts to targeted marketing for digital thieves. It can enable them to craft their cons in the language of specific audiences and regions, such as writing in different dialects of English.
In her example, criminals could use what she called the Matthew McConaughey style of Texas English to appeal to people in that state, referencing the actor known for his particular Southern drawl.
Digital criminals could also build “synthetic” identities from real ones and even manage to land remote jobs, Payton said, presenting another challenge for public agencies, tech providers and other organizations in their never-ending fight against fraud.
GENERATIVE AI AND WORKFORCE
Panelists at the conference, located just outside Washington, D.C., also talked about the intersection of AI and workforce, two of the main issues that occupy state CIOs, according to Meredith Ward, deputy executive director of NASCIO.
The jumping-off point for that discussion was a recent NASCIO and McKinsey & Co. report about how generative AI will impact state tech workers. For the most part, the report found that CIOs are optimistic about how AI can boost productivity and automation while improving services and data analysis.
“There is no fear from CIOs that robots will take over their jobs,” Ward said. AI could instead “free up [tech workers] from repetitive tasks,” one of the appeals of the emerging technology.
More specifically, procurement is among the tasks that AI could significantly improve, said Shawnzia Thomas, Georgia’s CIO and executive director of the Georgia Technology Authority. AI could reduce the procurement process from 18-36 months to six months, she said.
All that said, “the workforce is not ready for AI,” Thomas said.
It’s not from lack of enthusiasm or effort.
She said Georgia — which Ward called a “leader in this space” — has come up with an AI curriculum for tech staff, with workers “excited” about the technology. An AI innovation lab, as she called it, is crafting AI use cases.
“We are showing staff what AI can do for them,” Thomas said. “We are upskilling the staff we have.”
She also said she realizes that Georgia must keep boosting its AI game to attract younger tech professionals or graduates to the public sector. After all, she told attendees, younger people already are using AI and are gaining expertise relatively quickly.
In fact, Thomas touted efforts in Georgia to work with universities on AI and even conduct “road shows” for high school students.
The theory is that the public-sector tech professionals of the future might just need certifications and direct training instead of full, four-year, traditional college educations, which can be prohibitively expensive.
GENERATIVE AI AND PRIVACY
Yet another signal about the future of AI in the public sector came during a session about how generative AI will react to privacy concerns.
“Current laws have not caught up,” said Caterina Pañgilinan, the state chief privacy officer of Maryland.
Indeed, training of AI software promises to dredge up data that, under “record series” laws, should have been disposed of. That could, in turn, erode trust among citizens, many of whom are already suspicious of the data held by public agencies and their plans for AI.
Solving the privacy issue will require significant attention from state tech leaders in the coming months and years, according to Christopher Bramwell, Utah’s chief privacy officer.
“If your personal data is not ready for AI, you are not ready for AI,” he said.
