Regarding use cases related to fraud, Cotterill highlights the unemployment compensation aspect and appreciates the Chief of Staff and CIO of the Department of Workforce Development for seeding an AI-enabled workforce recommendation engine.
Through the system, an unemployment insurance claimant can opt for career recommendation based on work history and education level, says Cotterill. Moreover, he shares that the model has been trained on massive education and workforce datasets that MPH already had access to.
The challenge, according to Cotterill, lay in figuring out how to enable the right individuals to access the right data for the right purposes. He mentions working together with the team to train the model and creating an enabled environment separate from the underlying data.
Moving forward, Cotterill advises fellow professionals that winning friends by adding value is the way to work. While bureaucracy will always be there, it is imperative to think creatively and build relationships to articulate the why, how, and what. He adds that learning the federal rules is a must to materialize an idea.
When asked about getting legislative support, Cotterill acknowledges the constant support of the legislature towards MPH. With legislative funding of US$9 million a year, the MPH has attempted to build on that by contributing to agencies that have federal funding and creating platforms and products that benefit all.
Moving forward, Cotterill speaks about working on an AI policy to put guardrails to foster the adoption of trustworthy AI in Indiana. Furthermore, he mentions that the real value lies in partners telling the stories of organizational success and data.
Commenting on technology approaches to aggregate massive data, Cotterill mentions the human services vertical and the public safety vertical. He imparts that in those verticals, the recidivism rate can flare up, demanding the need to leverage from a tech perspective.
Cotterill mentions having a relational database management tool and cloud-based record linkage to ensure the accuracy of record linkage. He shares that although it is probabilistic, it has an aggregate accuracy as the products are generated. Further, the cloud perspective provides velocity to meet business user use cases, he adds.
Furthermore, Cotterill asserts the MPH has been a Microsoft user organization historically while having many other providers across the state government enterprise.
When asked about strategic objectives, he reiterates building an AI policy for executive agencies that adopt the NIST AI risk management framework as a maturity assessment.
Delving further, he affirms during the planning design development of an AI-enabled system MPH involves vendors and state agency partners to build the risk management framework. Additionally, a pre-deployment maturity assessment is carried out as the organization does not want its AI-enabled systems to perpetuate biases seen as negative.
The assessment deals with the cybersecurity risks associated with the system, respecting individual privacy and IP situation, says Cotterill. Also, the team has designated a privacy officer across 100 plus business units.
Cotterill informs that the MPH is funding the training of 65 privacy officers and helping them become certified information privacy professionals. He adds that the MPH wants the right individuals to see the right data for the right purposes and duration.
In conclusion, he states that the small team has managed to build privacy, data science, and governance functions within MPH. However, he notes that the business units need to take the regulatory policies forward, with their CDOs, CIOs, privacy officers, technical understanding, and asking the right questions.
CDO Magazine appreciates Ted Cotterill for sharing his success stories with our global community.
