As the use of artificial intelligence in everyday life grows, teachers, police departments and even state regulators are taking a cautious approach while finding ways the technology can be integrated into existing infrastructure.

Educators at Spruce Mountain Middle School in Jay are experimenting with using AI in a limited capacity, stepping forward, careful not to let it go too far. Cherry Poirier, a staff member at the Regional School Unit 73 school, has begun using AI as a classroom support tool while emphasizing caution and human oversight.

“I use it as a tool that assists me but it’s important to not let it do the thinking for me,” said Poirier, of Livermore Falls.

Poirier said she has used AI to create differentiated materials for students, particularly those in special education, but emphasized limits on student use.

“In our school, the students are not allowed to use it for their work as they still need to develop critical thinking and foundational skills,” she said.

Poirier also raised broader concerns about the technology’s long-term impacts.

“The consequences of using it too much can not only be detrimental to the environment but also to our cognitive abilities,” Poirier said, adding that “it’s super important to always have a human in the loop.”

Law enforcement agencies are also exploring limited uses of artificial intelligence while emphasizing caution and oversight.

“In present-day society, most of us are using what could be defined as AI tools in our day-to-day lives,” Farmington Police Chief Kenneth Charles said. “Spell and grammar check is a great example. Navigation tools, such as Google Maps or Waze, online streaming of music or movies, even social media. Even e-commerce uses AI.”

Charles said one current application within the department involves processing public records requests under Maine’s Freedom of Access Act.

“Redaction tools are used to search for and redact restricted information,” he said, noting that the technology is helpful “but not fail-safe.”

He added that the department is developing a formal policy and stressed that employees remain responsible for ensuring accuracy and protecting sensitive data.

“While it may be a useful tool, the employee is still responsible for their work product,” Charles said.

As artificial intelligence tools increasingly appear in classrooms, police departments and workplaces across Maine, state government is taking a cautious approach, allowing limited use while placing strict restrictions on how employees interact with the technology.

The Maine Office of Information Technology has issued a Generative Artificial Intelligence Policy outlining how executive branch employees may use tools such as chatbots and automated content systems. The policy allows some limited use but requires careful oversight and strong protections for government data.

Generative AI systems are technologies capable of creating new content such as text, images, code or other materials that resemble human-generated work. The tools could help government employees complete routine tasks more efficiently, but they also present security and reliability risks that require clear safeguards.

One of the policy’s central restrictions is that employees may not enter confidential or restricted government information into generative AI systems unless the technology has been formally approved through the state’s technology review process. The rule is intended to prevent sensitive data from being transmitted to external systems that may store or process information outside the state’s secure networks.

The policy also requires human review of any AI-generated work used in government operations. Because generative AI tools can sometimes produce inaccurate or fabricated information, employees must verify the accuracy of the output before relying on it in their work.

RISE OF AI

Maine’s cautious approach developed after the rapid rise of tools such as ChatGPT in 2023. That year, state cybersecurity staff temporarily barred executive branch agencies from using generative AI systems while they studied the security, legal and privacy implications of the technology.

“In June 2023, Maine’s Office of Information Technology imposed a six-month moratorium on the use of Generative AI … in executive branch agencies,” said Sharon Huntley, director of communications for the Department of Administrative and Financial Services. “This pause, ultimately extended a further three months, gave the state time to study the new technology’s implications.”

Huntley said the state conducted risk assessments focused on “security and privacy threats,” while also examining “potential bias and ethical issues” and reviewing federal guidance before moving forward with a policy framework.

State cybersecurity leaders warned that AI systems can produce unpredictable results and could expose government data if used improperly. The technology also raises concerns about misinformation, algorithmic bias and potential legal issues involving copyrighted material.

Despite those concerns, the state’s policy acknowledges that artificial intelligence may eventually provide important benefits for government agencies.

AI tools can analyze large volumes of information, summarize documents and assist with repetitive administrative tasks. Supporters say these capabilities could help public employees manage workloads and improve the efficiency of government services.

By early 2024, that review led to the creation of guiding principles and an acceptable use policy, Huntley said, “aligning Maine’s approach with emerging best practices.”

Recognizing the broader implications of the technology, Gov. Janet Mills signed an executive order in 2024 creating a statewide artificial intelligence task force. The group was charged with studying the technology’s potential effects on Maine’s economy, workforce and public institutions.

The task force is examining both opportunities and risks associated with AI and has been developing recommendations to guide future policy decisions regarding sectors like healthcare, education, manufacturing and government operations across the state.

Artificial intelligence is already appearing in some areas of Maine’s economy. Businesses and research institutions have begun using AI tools to analyze data, improve manufacturing processes and support scientific research.

State policymakers say the challenge is determining how government agencies can safely adopt similar technologies without exposing sensitive information or relying on systems that may produce unreliable results.

Huntley said agencies seeking to use new AI tools must go through a formal review process with the MaineIT Architecture and Policy Team.

“A key concern … is determining the sensitivity of the data that might be at risk,” she said, noting the state uses a classification system based on federal cybersecurity standards to limit AI use to low-risk information.

She added that the policy requires human oversight in all cases.

“By policy, we do not allow a GenAI product to be employed without … the ‘Human in the Loop’ compensating control.”

MaineIT also evaluates whether AI systems can produce incorrect results.

“It’s important MaineIT makes sure the product’s own internal controls guard against these failures,” she said, referring to so-called hallucinations.

For now, Maine’s policy treats artificial intelligence as a tool to be studied and carefully tested rather than widely deployed in government operations.

“MaineIT and State of Maine use of Generative AI tools will continue to evolve and expand while carefully ensuring the safety and security of the data and resources that Maine people entrust in our care,” Huntley said.