rioritize AI Use Cases for Education

Address the potential of AI to transform education.

Analyst perspective

AI is rapidly entering the education space, and CIOs need to be prepared to take advantage of its potential benefits. An AI use case library for education is designed to support that need.

The strategic priority of AI determines the institution's approach. For example, institutions focused on institutional growth and sustainability may use AI to personalize learning, optimize course offerings, and identify high-potential students. Institutions focused on operational excellence may use AI to automate tasks, improve efficiency, and reduce costs. Institutions focused on instructional and research value may use AI to create personalized learning experiences, provide real-time feedback, and discover new knowledge.

The introduction of AI can be contentious, and the risks should be considered carefully. AI can have biases that directly thwart the mission of the institution. It is also a new technology, and its promise still outweighs its results.

Finally, many IT shops need to develop capabilities to support AI, and a clear strategy is necessary to plan for this development.

Mark Maby
Research Director for Education
Info-Tech Research Group

Executive summary

Your Challenge	Common Obstacles	Info-Tech's Approach
Generative AI has increased the education sector's interest, concerns, and expectations for artificial intelligence. They will turn to IT for guidance on how AI can serve their institutions. Adopting new technology requires a strategic approach and alignment between IT and the business. AI technologies are typically significant investments. A smaller organization with limited resources will need to make a comprehensive business case to justify the investment.	Educational institutions are concerned about the risks, compliance, regulations, and policies of AI and ML. Institutions have a limited understanding of how AI can impact them and how to get started with prioritization. Determining relevant use cases for the education sector can be difficult and time-consuming.	Discover and comprehend the relevant use cases that can address organizational challenges. Begin the AI journey by identifying and prioritizing use cases for their departmental units through the use case analysis tool. Leverage the output to gain executive buy-in. Determine the problems with the greatest-value solutions and meet institutional needs to implement AI responsibly.

Info-Tech Insight

The approach to artificial intelligence should be strategic and responsible, with a clear understanding of the relevant use cases and benefits and a plan to address the challenges of implementation and ongoing use. Educational institutions that invest in AI will foster innovation to improve operational efficiency, student and faculty experiences, and data-driven decision-making.

AI adoption in the education space is driven by learner outcomes

Top Three Barriers to AI Adoption

Lack of talent with AI skills	53%
Under-resourcing for Al	50%
Lack of clear strategy	47%

Info-Tech Insight

The responses on this page reflect the perspectives of leadership in the technology sector.

Primary motivations for AI adoption are enhancing learner outcomes and cost efficiency:

• Make instruction more adaptive and personalized to the needs of the student.
• Make processes more efficient, not the least for teachers.

However, both AI technologies and skill development come with investment requirements:

• Most importantly, this includes the integration of data with the AI and training and recruiting staff to effectively use AI tools.

These opportunities and barriers highlight the necessity of a clear AI strategy:

• Where AI initiatives are aligned with institutional goals.
• Where use cases are specified for relevance to the institution.

AI will benefit educators the most

Top AI opportunities and concerns identified by educators

Opportunities	Pct.	Concerns	Pct.
Boosts efficiency	73%	Potential for cheating	38%
Thought starter/Idea generator/springboard	68%	Potential to stifle creativity	38%
Information at fingertips	53%	Concern about focus on product over process	36%
Automate mundane tasks	53%	Incorrect or fabricated results	27%
Personalized teaching/24-hour TA access	31%	Equity and access	38%

(Ghimire, et al., 2024)

Seven new national AI research institutes

National Science Foundation invested $140 million in AI research.

Two of the seven focus on researching AI implications on education.

(NSF News, 2023)

38%

Percentage of students aged 12-18 who admit to using ChatGPT for an assignment without their teacher's knowledge.

(Common Sense Media, 2023)

Info-Tech Insight

The data on this page were provided by educators.

By coincidence, that potential and admission of using AI to cheat are both surveyed at 38%.

While AI is promoted for personalized teaching, the main benefits are for supporting educators in their processes and less about the benefits for learning.

The strategic priority of AI determines the institution's approach

Dedicated Team for AI and Digital Products
The institutional strategy has identified AI as a priority and created a dedicated team with functions for AI engineering, systems architecture, business analysis, and software development.

Prioritized Use Cases
The institutional strategy has identified specific, high-impact use cases involving AI. These likely require both outsourced development of the solution and resources to maintain the operations of the technology.

Attentive Adoption
Commercial-off-the-shelf (COTS) AI technologies and features are becoming commonplace. Policies are in place to address technological and other institutional risks due to their adoption.

Uncontrolled Proliferation
AI products and features are becoming common with little oversight.

Most institutions will focus their approach at the levels of Attentive Adoption or Prioritized Use Cases.

Dedicated team oversees AI and other digital products. Very few institutions are here.

Specific, strategically important use cases are prioritized.

Policies are developed to address AI and its adoption.

AI use proliferates among shadow IT with little oversight.

Generative AI is an innovation in machine learning

Generative AI (Gen AI)
A form of ML whereby, in response to prompts, a Gen AI platform can generate new outputs based on the data it has been trained on. Depending on its foundational model, a Gen AI platform will provide different modalities and thereby use case applications.

Machine learning (ML)
An approach to implementing AI whereby the AI system is instructed to search for patterns in a dataset and make predictions based on that set. In this way, the system learns to provide accurate content over time (think of Google's search recommendations).

Artificial intelligence (AI)
A field of computer science that focuses on building systems to imitate human behavior. Not all AI systems have learning behavior; many systems operate on preset rules, such as customer service chatbots.

Info-Tech Insight

Many vendors have jumped on "Gen AI" as the latest marketing buzzword. When vendors proclaim to offer Gen AI functionality, pin down what exactly is generative about it. The solution must be able to generate new outputs – not merely predictive outputs.

Other technologies involved with AI use cases

Adaptive learning algorithms:

• Algorithms that adjust their behaviors based on the learner's performance
• Personalized learning, adaptive assessment

Computer vision:

• The extraction of meaning from digital images or videos
• Self-driving cars, facial recognition, medical imaging

Game engines:

• Software to create and run video games
• Gamification in instructional software

Natural language processing (NLP):

• The interaction between computers and human (natural) languages
• Machine translation, text analysis, speech recognition

Natural language generation (NLG):

• NLP to create human-like text
• Chatbots, virtual assistants

Machine translation (MT):

• NLP to translate text from one language to another

Personalization algorithms:

• Algorithms that tailor their output to the individual user
• Product recommendations, news feeds

Predictive analytics:

• The use of statistical models to predict future outcomes
• Fraud detection, machine failure

Text mining:

• The extraction of knowledge from text documents
• Sentiment analysis, topic modeling, spam filtering

Download the Get Started with Artificial Intelligence blueprint to learn more

Artificial intelligence performs tasks mimicking human intelligence. AI is a combination of data-driven technologies that include tools such as machine learning, technology that learns through experience and by problem-solving.

The discussion of AI can often become too broad because the term often refers to multiple technologies. To the left you'll find specific technologies used in conjunction with machine learning and generative AI.

This report includes a use case library for education. These different technologies are specified in the library to clarify what type of AI the use case is referring to.

Consider the risks of AI

There are more than the usual number of risks with AI technology.

MITIGATION FACTORS

Trust

Transparency: Can the system explain its decision in an understandable way to users?

Control: Are there procedures for detecting and responding to errors, as well as mechanisms for human oversight?

Trainable: Can the AI system be retrained using a diverse dataset to identify and remove bias from the data?

Continuous improvement

Institutions should continuously monitor the use of AI-enabled technologies to ensure they are meeting the needs of their users and being used safely and ethically.

RISKS

Bias

Many large language models (LLM) are trained on data from the internet, adopting its biases as well as those of their trainers.

Accountability

Ultimately, the institution will be accountable for the decisions of the AI tool, including the issues around copyright. The systems are often opaque, thwarting mitigation techniques.

Technology

Accuracy: The models are often inaccurate and have "hallucinations," where responses are not based on observation.

Shadow IT: There is likely uncontrolled implementation and use of AI among constituents.

Vendors: AI is a new landscape, and the suppliers lack maturity.

Privacy and security

Concerns around data privacy and security are both typical of technology and novel to the strangeness of AI.

An AI use case library for education

Leverage best-in-class digital use cases to build strong implementation roadmaps and maximize value creation.

An AI use case is a technology incorporating artificial intelligence and applied to a specific capability within a given industry to create value.

Consider the factors presented here when assessing the value of a use case.

Technology
What base technology is applied to deliver the use case?

Benefits
What value does the use case provide to the organization.

Industry
A use case often applies to both higher education and K-12, but not always.

Value Streams
Value streams are specific to each industry. They organize the organization's core capabilities according to the value it delivers value to its constituents.

Risks
Consider potential issues when adopting the technology.

Feasibility
How feasible is implementation of the use case, based on prevalence in the education sector?

Capabilities
Capabilities define how the organization functions through the interaction of its people, processes, and technology.

Opportunities for using generative AI in cybersecurity

Opportunity 1: Security incident simulation

1. Incident initiation: Create a cyberattack scenario, like an elaborate phishing attack, within a described context that matches your organization.
2. AI-driven attack dynamics: The AI is prompted to be adaptive and evolve its attack patterns, changing tactics in response to the actions taken by the incident response team, mimicking the behavior of real-world cyber threats.
3. Role assignments and communication: Participants are assigned specific roles within the incident response framework, such as Incident Commander or Communications Lead
4. Decision-making and escalation: Throughout the exercise, teams must make critical decisions under pressure, such as whether to shut down systems, engage law enforcement, or communicate with stakeholders.
5. Real-time feedback and adaptation: The AI system provides real-time feedback, including simulated media coverage, stakeholder reactions, and the unfolding impact of the cyberattack.
6. Post-incident analysis: After the simulation, teams review their actions, discuss what worked well, and identify areas for improvement. The AI can also generate detailed reports summarizing the incident timeline, key decisions, and their outcomes.

Generative AI can simulate cyberattacks for incident response training. Effectively, the AI is prompted to be the "game master" and create a tabletop exercise for incident preparedness.

Such a simulation can enhance the readiness and effectiveness of cybersecurity teams by mimicking realistic cyber incidents and their complex dynamics.

See "Prompting for cyber incident response practice- a generative AI example"

Opportunities for using generative AI in cybersecurity

Opportunity 2: Security incident communication

1. Incident communication setup: Define the types of information that need to be conveyed to stakeholders during a security incident.
2. Data preparation and input structuring: Organize messy and unstructured data (e.g. text, logs, images, links, stats, timelines, code snippets) into a structured format using self-explanatory tags like to align the data with incident communication templates.
3. Prompt engineering: First create simple prompts to instruct the LLM to summarize the incident facts. Then refine prompts by adding guidelines for clarity and key point coverage. Use tags to highlight important content and guide the LLM. Include examples of good incident summaries as models.
4. AI-driven summary generation: Use LLMs to generate incident summaries, ensuring they cover all key points and follow writing best practices (e.g. neutral tone, active voice, minimized acronyms).
5. Integrate AI into workflow: Integrate a "Generate Summary" button in the incident management UI that allows a human user to accept, modify, or discard the generated response.

Leveraging generative AI can enhance the efficiency and effectiveness of security incident response processes. These steps describe how to implement a system that uses LLMs to generate high-quality summaries and communications, ensuring timely and accurate information dissemination.

See "Accelerating incident response using generative AI"

Measure the value of this document

Document your objective

Highlight best-in-class use cases to spur the initiative-planning and ideation process.

Measure your success against that objective

There are multiple qualitative and quantitative, direct and indirect metrics by which you can measure the progress of your initiative pipeline's development. Some examples are:

• Increased initiative pipeline value.
• Number of capabilities impacted by initiative pipeline.
• Enhanced understanding of the initiatives' impact aligned to the organization's capability map.
• Better understanding of which sources of value are being addressed or under-addressed in the organization's initiative pipeline.

Examples:

Expected Outcome	Project Metrics
Increase throughput of use cases	Increase in number of use cases piloted and implemented compared to last year Time from idea submission to scale
Select valuable use cases	Customer satisfaction Business satisfaction Decreased costs Increased revenue

See Identify and Select Pilot AI Use Cases in the Artificial Intelligence Research Center for more details

Leverage the higher education capability map to identify candidate opportunities and initiatives

Business capability map defined…

In business architecture, the primary view of an organization is known as a business capability map.

Business capability defines what a business does to enable value creation, rather than how. Business capabilities:

• Represent stable business functions.
• Are unique and independent of each other.
• Will typically have a defined business outcome.

A business capability map provides details that help the business architecture practitioner direct attention to a specific area of the business for further assessment.

Leverage the K-12 education capability map to identify candidate opportunities and initiatives

Business capability map defined…

In business architecture, the primary view of an organization is known as a business capability map.

Business capability defines what a business does to enable value creation, rather than how. Business capabilities:

• Represent stable business functions.
• Are unique and independent of each other.
• Will typically have a defined business outcome.

A business capability map provides details that help the business architecture practitioner direct attention to a specific area of the business for further assessment.

Capabilities tree

Level 1: Value streams
Core components of an organization's value chain or support structure

Level 2: Capabilities
The top-level activities that your organization performs to ultimately deliver a product/service

Level 3: Subcapabilities
The subactivities, or jobs to be done, performed within an overarching capability

Download the Higher Education Industry Business Reference Architecture Template

Download the K-12 Education Industry Business Reference Architecture Template

Use cases apply to a specific level 3 capability within the industry value stream.

Leverage value drivers for education to align with institutional strategy

Institutional growth and sustainability	Drives sustainable growth, diversifies methods of generating revenue and decreasing costs, and increases student/institutional market reach.
Operational excellence	Provides transparency in the flow of value to the students and faculty, empowers administrative staff, and promotes teamwork.
Instructional and research value	Enhances the experience of students and faculty in their studies. It also supports the funding, development, and dissemination of academic and applied research.
Risk and resilience	Mitigates and withstands rapid changes across the IT landscape, secures student and academic information while protecting personal and institutional information, and easily integrates with current technologies, projects, and strategies.
Brand impact, community engagement, and social responsibility	Differentiates the institution from competitors to external communities while strengthening its position on social responsibility.

Value drivers are factors that impact the success, effectiveness, and overall value of educational institutions or programs.

The five factors listed here are used to organize the use cases presented in this report.

The quality of educational outcomes is the ultimate driver of value; however, the institution is also an organization of people that must be self-sustaining and functional. These drivers are presented as the motivating factors for any strategic initiative within education.

There are distinctions between K-12 education and higher education, as well as between publicly-funded and private institutions. With some small modification, these drivers should be broadly applicable to any institution of education.

Despite its benefits, AI may not align with the mission of education

The strategic value of AI is subordinate to the larger attitude of AI within the educational community.

AI can advance strategic priorities

Institutional growth through enhanced marketing

Operational excellence by reducing the burden of repetitive activities

Instructional value by tailoring instruction to the individual student

Risk resilience through the automation of cyber-threat detection

Community engagement through increased responsiveness

The mission of education is at odds with AI

Reduced opportunities for human contact and professional judgement

Students prevented from learning essential skills such as academic researching and evaluation

Potential for discrimination, bias, and privacy violations

Staff threatened with displacement and find the technology intrusive

Unacceptable to the local purpose, culture, and community

Info-Tech Insight

IT leadership should involve themselves in the debate around AI at their institution to identify cultural restrictions.

What is an AI use case?

An AI use case is a technology or combination of technologies applied to a specific capability (e.g. job to be done) within a given industry/function to create value.

Use case

Capabilities
The activities, or jobs to be done, that your organization performs to ultimately deliver a product/service

Technology
The base technology that enables value-creating performance gains

Industry or function
The relevant industry or function (many use cases will apply across multiple industries/functions)

The AI use case library

What is it?

A use case represents a technology or combination of technologies applied to a capability within a given industry or function that drives value. The AI use case library is a nonexhaustive list of Gen AI/AI/ML use cases that can be organized by industry/function, capability, or technology. The organizing principle in this document is by industry/function.

Why is it important?

In the context of a digital transformation, the Gen AI/AI/ML use case library:

• Identifies potential sources of value to analyze in a top-down opportunity assessment.
• Jumpstarts the idea generation process during the initiative development phase. Use cases are the foundational building blocks of the initiatives that ultimately deliver value to the business.

Leverage Info-Tech's Identify and Select Pilot AI Use Cases blueprint

Use this research to gather a longlist of potential AI use cases.

The present research provides a longlist of use cases that integrate with Info-Tech's methodology on selecting AI use cases.

Identify and Select Pilot AI Use Cases supports:

• Creating a team to develop and select use cases.
• Assessing the value of the various use cases.
• Developing a pilot project

Each of the use case slides follows the same format

1 Leading AI use cases for institutional growth and sustainability

No.	Use Case	Industry	H.E. Value Stream	K-12 Value Stream	Description	Technology	Benefits	Risks	Feasibility
1.01	Admissions chatbot	Higher Ed	Admission	--	An AI-powered chatbot assists prospective students with admissions-related queries.	NLP, Gen AI	Efficient admissions support, improved student experience	Lack of human interaction, potential bias in responses	◑
1.02	Candidate review	Higher Ed	Admission	--	By analyzing applications, AI can predict which students are most likely to be a good fit for the institution based on test scores and extracurricular activities.	NLP, Gen AI	Efficient candidate selection process, improved student success	Potential bias in selection criteria, privacy concerns	◔
1.03	Financial aid support	Higher Ed	Student enrollment	--	Design aid packages and make instant decisions for at-risk students.	NLP, Gen AI	Improved financial aid allocation, increased access to education	Potential bias in decision-making, privacy concerns	◑
1.04	Marketing	Both	Recruitment	Student enrollment	Analyzing data from prospective students' social media profiles and websites, the system can create personalized marketing materials tailored to the interests and goals of each student.	NLP, ML, text mining	Targeted marketing, improved student recruitment	Privacy concerns, potential bias in targeting	◑

2 Leading AI use cases for operational excellence

No.	Use Case	Industry	H.E. Value Stream	K-12 Value Stream	Description	Technology	Benefits	Risks	Feasibility
2.01	Administrative tasks	Both	Supporting capabilities	Supporting capabilities	Automate repetitive tasks, leading to increased efficiency and workflow support.	ML, Gen AI	Time-saving administrative processes, increased productivity	Reliance on automation, potential system errors	◔
2.02	Advising	Both	Student support services	Supporting capabilities	A virtual advising assistant resolves electronic queries from students by automating certain tasks and enabling live advisors for complex tasks.	NLP, ML, Gen AI	Efficient advising process, improved student support	Reliance on automation, potential bias in advice	◑
2.03	Automated scheduling	Both	Enrollment	Manage student enrollment	Generate a full and workable schedule for the upcoming semester.	ML	Saves time and effort for administrators, reduces scheduling conflicts	May overlook specific constraints or preferences, potential disruptions caused by errors in automated process	◕
2.04	Energy management	Both	Supporting capabilities	Supporting capabilities	Energy consumption is monitored and adjusted through thermostat settings, lighting levels, and ventilation.	ML	Energy efficiency, cost savings	Reliance on automation, potential system errors	◕

Virtual agent reduces workload and improves remote support

Rapid Response Virtual Agent provides 24/7 support for students and parents.

INDUSTRY
K-12 Education

SOURCE
Quantiphi, 2023

Challenge	Solution	Results
The Upper Grand District School Board was inundated with an unprecedented number of inquiries for reliable information pertaining to online safety and technical support. The school board was struggling to keep up with the demand, and the workload was putting a strain on administrative staff.	The board contracted Quantiphi to develop a Rapid Response Virtual Agent to provide quality, round-the-clock support. The bot was integrated with the school board's existing website and a content management system, Google Sheets, for easy updates and modification of intent responses. The agent handles 1,000 queries per day and is trained to handle 20 distinct FAQs.	The virtual agent has reduced the workload of administrative staff by 50%. The agent has handled over 3,000 queries in the first week of implementation. The agent has a 92% accuracy rate in resolving queries.

2 Leading AI use cases for operational excellence

No.	Use Case	Industry	H.E. Value Stream	K-12 Value Stream	Description	Technology	Benefits	Risks	Feasibility
2.05	Facilities management	Both	Supporting capabilities	Supporting capabilities	AI can provide insights into facilities performance (e.g. identifying areas at risk of wear and tear). It can also automate tasks such as scheduling maintenance and ordering supplies.	ML	Improved facilities maintenance, cost savings	Reliance on automation, potential system errors	◕
2.06	Forecast of class demand	Higher Ed	Enrollment	--	Generate a projected number of sections per subject to offer or open for the upcoming term.	ML	Optimized resource allocation, reduced under- or over-enrollment.	Inaccuracies may lead to scheduling issues and resource wastage.	◕
2.07	Processing transcripts	Higher Ed	Course completion and graduation	--	Optical character recognition (OCR) extract data from physical transcripts and AI to identify errors and inconsistencies.	OCR, ML	Automated transcript processing, improved data accuracy	OCR errors, potential bias in data processing	◔
2.08	Recruitment	Both	Supporting capabilities	Supporting capabilities	Identify potential applicants who meet the minimum qualifications for open positions.	ML, Gen AI	Efficient recruitment process, improved candidate selection	Potential bias in selection criteria, privacy concerns	◔
2.09	Contract review	Both	Supporting capabilities	Supporting capabilities	Analyze contracts for specific clauses and changes to previous contracts to speed up review processes.	Gen AI	Time-saving and accuracy	Risk of missing key information	◑

3 Leading AI use cases for instructional value

No.	Use Case	Industry	H.E. Value Stream	K-12 Value Stream	Description	Technology	Benefits	Risks	Feasibility
3.01	Adaptive learning	Both	Teaching and learning	Deliver instruction	Educational software tailors the learning experience to each individual student by tracking student progress and adjusting the difficulty of the material accordingly.	Gen AI, adaptive learning algorithms	Improved learning experience, improved student outcomes, increased engagement	Privacy concerns, potential bias in content delivery	◑
3.02	AI teaching assistant	Both	Teaching and learning	Deliver instruction	AI-powered systems support teachers and students with feedback and personalized learning plans.	NLP, ML, computer vision	Time-saving for teachers, personalized support for students.	Potential bias and tracking without consent	◑
3.03	Grading assignments	Both	Teaching and learning	Assess student achievement	Grade both written and typed responses using a key or rubric and provide feedback.	ML, Gen AI	Efficient grading process, timely feedback for students	Reliance on automated grading, potential bias in grading	◑
3.04	Identifying at-risk students	Both	Course completion and graduation	Deliver instruction	Identify students who are struggling academically, have attendance problems, or are not engaged in their studies.	NLP, ML, Gen AI, predictive analytics	Early intervention for at-risk students, improved student retention	False positives/ negatives, privacy concerns	◕
3.05	Materials creation	Both	Teaching and learning	Develop curriculum	Leverage generative AI in the process of creating both formative and summative assessments.	Gen AI	Efficient curriculum development, personalized learning materials	Quality control of generated materials, potential bias in generated content	◑
3.06	Personalized learning	Both	Teaching and learning	Deliver instruction	This form of adaptive learning allows the student to customize the tool to suit their needs.	NLP, ML, Gen AI, personalization algorithms	Tailored learning experience, improved student engagement	Privacy concerns, potential bias in content delivery	◑
3.07	Plagiarism detection	Both	Teaching and learning	Assess student achievement	AI identifies instances of plagiarism, helping students avoid it and teachers detect it.	NLP, ML, Gen AI	Promotion of academic integrity, time-saving for teachers	Inaccuracy, unfair targeting of students	◕
3.08	Proctoring examinations	Both	Teaching and learning	Assess student achievement	AI monitors students during online exams to reduce cheating and ensure fairness. This can combine AI technologies with video monitoring and browser restrictions.	Computer vision, ML	Reduced cheating and increased fairness during online exams	Invasion of privacy, ethical implications	◕

AI in the classroom

A number of Gen AI tools for dedicated classroom use address the challenge of using public LLMs like ChatGPT in the classroom.

Public Gen AI tools are inappropriate for classroom use:

• Public tools are designed for users 13-years or older.
• Educators want to ensure safe, suitable content for younger students.
• Public models may not meet privacy requirements such as COPPA, PIPEDA or the Children's Code, depending on the jurisdiction.

Potential solutions:

• Red teaming
  • Ethical hackers adapt LLM models for safety and appropriateness.
  • This is cost-prohibitive for most districts.
• Dedicated applications
  • Use child-friendly applications (e.g. Khanmigo, Eduaide.Ai, SchoolAI, Padlet, aiEDU, Magic School, TeachMateAI, LessonPlans.ai).
  • These have built-in safeguards for content and privacy and reduce the learning curve for educators.

AI platforms for classroom instruction

AI tools used in a classroom should offer the standard data privacy safeguards expected of any classroom software.

Lesson planning

• Tools for generating unit and lesson plans
• Specialized tools for detailed plans, such as rubric generators, academic content generators, and misconception alert systems

Activity creation

• Support for active learning environments, such as math practice reviews, science lab generators, and quiz format diagnostics
• Facilitation of various teaching strategies, such as inquiry-based, project-based, and game-based learning

Student feedback and support

• Tailored support for individual needs, IEP (individual education plan) suggestions, text scaffolding for reading levels, and accommodation suggestions
• Detailed feedback on student work that identifies strengths and improvement areas

Writing and communication

• Assistance in writing and communication tasks, text rewriting and proofreading tools, professional email generators, and class newsletter creators
• Efficient communication with parents/guardians, structured templates, and automated response generators

How are faculty dealing with the intervention of generative AI?

Plagiarism detection tools are by no means foolproof. Forward-thinking faculty are embracing the new world and integrating generative AI into their classes.

Writing research before generative AI	Writing research after generative AI
Identifying research interest Identifying source material Reading and writing research notes Writing paper with documentation Completing works cited or bibliography	Identifying research interest Writing paper with generative AI Checking for hallucinations Identifying academic knowledge Identifying source material Documenting sources Completing works cited or bibliography

Source: Kassorla, "Teaching with GAI in Mind," 2023

• Embrace generative AI as a tool for teaching and learning.
• Learn as much as you can about generative AI and its capabilities.
• Teach with the benefits of generative AI or warn of its limitations depending on the purpose of your class.
• Make a clear syllabus statement about your generative AI policy.
• Use project-based learning and alternative assessment methods.

Info-Tech Insight

Higher education may want to invest in curriculum management as methods of assessment change.

AI detection tools require an effective academic honesty policy

These tools vary in effectiveness, with none being foolproof.

Numerous AI detection tools have appeared. They can give you a confidence measure but not an absolute decision. False positives are likely and problematic. Examples of these tools include:

• Originality.ai
• Crossplag
• Copyleaks
• Turnitin
• Content at Scale

There is a cat and mouse game between the AI, which strives to be natural, and the detection tool, which detects artificial language.

Educators need a strong academic honesty policy that is transparent about how these tools are used.

4 Leading AI use cases for academic research value

No.	Use Case	Industry	H.E. Value Stream	K-12 Value Stream	Description	Technology	Benefits	Risks	Feasibility
4.01	Automated writing	Higher Ed	Research	--	AI generates essays automatically, saving time for researchers.	Gen AI	Time and effort savings for academics	Risk of plagiarism, misrepresentation and peer-review issues	◑
4.02	Data analysis	Higher Ed	Research	--	Gen AI can assist preliminary data analysis tasks for academic research, such as data cleaning, modeling, and visualization.	NLP, ML, Gen AI, statistics	Efficient data analysis, improved research outcomes	Data quality issues, potential bias in analysis	◕
4.03	Document AI	Higher Ed	Research	--	AI technology can process and analyze documents for academic research purposes.	NLP, ML, Gen AI, OCR	Efficient document analysis, improved research outcomes	Accuracy of analysis, potential bias in interpretation	◔
4.04	Grant application	Both	Research	Supporting capabilities	Draft initial versions of grant applications for academic researchers.	NLP, Gen AI	Time-saving	Risks of plagiarism, lack of originality	◔
4.05	Literature review	Higher Ed	Research	--	A search engine can find relevant literature, use generative AI to summarize it, and identify possible trends.	NLP, Gen AI, text mining	Time-saving literature review process, identification of research trends	Quality of generated summaries, potential bias in content	◑
4.06	Peer review	Higher Ed	Research	--	AI can be trained on a large corpus of academic papers and their corresponding peer reviews to generate new reviews based on the content of a given paper.	NLP, Gen AI, text mining	Quicker manuscript evaluation, draft reviewer commentary	AI biases in review, confidentiality risks	◔
4.07	Synthetic data	Higher Ed	Research	--	Create synthetic data for research and development.	NLP, ML, Gen AI, computer graphics	Accelerated research and development, reduced need for time-consuming data collection and cleaning processes	The quality and representation of synthetic data may not fully reflect real-world data.	◕

Model 1: University of Michigan

Leveraging AI: University of Michigan's approach to building closed generative AI tools

INDUSTRY
Higher Education

SOURCE
Educause, 2024

Challenge	Solution	Results
Leadership at U-M recognized the importance of generative AI to its mission as a leading research institution. However, the public tools available raised policy issues for the university: Privacy issues with data sharing Compliance with web accessibility guidelines Costs of tools increasing inequality among students The university required AI tools that ensured accessibility and inclusivity for all community members.	The university decided to develop internal AI tools that kept queries private. U-M's ITS staff would update and tune the language model. This would ensure data security by avoiding continual training on user data. The ITS team combined vendor resources and internal development to create the tools. They licensed Microsoft's LLM for use within a secured Michigan environment.	The team created three tools: U-M GPT: A ChatGPT-like tool for answering questions, producing content, and making recommendations U-M Maizey: A no-code platform allowing users to create customized AI chat programs with their datasets. U-M GPT Toolkit: A tool for developers to build, train, and host their own AI models.

Model 2: University of Florida

AI integration in education: University of Florida's strategic approach

INDUSTRY
Higher Education

SOURCE
NVIDIA, 2024

Challenge	Solution	Results
Growing demand for AI skills: National increase in AI skill demand Rapid evolution of AI and cybersecurity fields High need for AI expertise across various sectors Skill gaps in graduates: Lack of practical AI experience in many graduates Projected global tech talent shortage Educational inequality: Limited AI resources in many colleges Funding disparities between public and private institutions	Partnership with NVIDIA: Installation of HiPerGator AI supercomputer Development of a clinical language model Funding from multiple sources: $50 million from NVIDIA and its founder $20 million from university colleges $130 million from government of Florida for a data center Faculty: Hired 100 AI-focused faculty Upskilling existing faculty through NVIDIA's Deep Learning Institute	Education initiatives: Specific courses (e.g. AI in Agricultural and Life Sciences, AI in Built Environment, Al in Media and Society) Nine-credit AI certificate for undergraduates Microcredentials in AI skills Initial research projects: Using AI to achieve state goals, national and global priorities Tracking of Florida land-use patterns Identification of at-risk students Reading and analyzing clinical notes stored in electronic health records.

Model 3: University of California at San Diego

An existing supercomputer makes in-house development the cheaper route to generative AI.

INDUSTRY
Higher Education

SOURCE
Baytas, 2023

Challenge	Solution	Results
UC San Diego had a growing reliance on commercial generative AI for administrative tasks. However, the public tools raised policy issues for the university: Ensuring data security Fine-tuning the model on their own data Managing community expectations of these tools Technical challenges in integrating AI with existing systems	The university was able to leverage their supercomputer to develop TritonGPT. TritonGPT is a suite of AI assistants that is powered by Meta Llama 3, an open-source, dialogue-optimized LLM. They developed a team to fine-tune the model who incorporated crowd-sourced feedback from early users using thumbs up/thumbs down responses and comments.	Full deployment to 37,000 faculty and staff within six months Substantial reduction in time spent on administrative tasks (e.g. job descriptions, summarizing documents) Ensured data privacy and security through on-premises hosting; no third-party data sharing Projected 40% cheaper over three years compared to cloud AI services

Key takeaways for higher education

Florida	UC San Diego	Michigan
The most comprehensive AI initiative in the current higher education landscape Largely beyond the scope of most universities without the existing infrastructure Highlights the benefits of a strong industry partnership as well as the need for a concerted funding effort Provides a useful curriculum model for AI that other colleges and universities can adapt and build upon	Provides an example of how a university can develop the necessary infrastructure over time Strongly positioned to develop AI skills among their students Two overarching goals for AI in university: Leverage benefits of AI to improve efficiency and competitiveness. Support the development of AI skills to prepare students for the workplace.	The most reproducible model of the three Leveraged an existing relationship with Microsoft Azure Provides private tools to faculty, students, and staff to protect institutional data (but did so before identifying use cases) May encourage other institutions to adapt the strategy to specific use cases due to low cost

5 Leading AI use cases for risk and resilience

No.	Use Case	Industry	H.E. Value Stream	K-12 Value Stream	Description	Technology	Benefits	Risks	Feasibility
5.01	Admissions fraud	Higher Ed	Admission	--	AI can be used to detect fraudulent applications by analyzing data points such as test scores, extracurricular activities, and personal essays to identify patterns of inconsistency in an application.	ML, data analytics, Gen AI	Enhanced application screening, reduced fraud	False positives/negative, privacy concerns	◔
5.02	Cyber-threat detection	Both	Supporting capabilities	Supporting capabilities	Use threat hunting and network monitoring tools to monitor all connected devices.	ML, data analytics	Enhanced cybersecurity safeguarding sensitive data and networks	False positives may lead to unnecessary disruptions and strain on resources.	◑
5.03	School security	Both	Supporting capabilities	Supporting capabilities	AI improves school security through monitoring and threat identification.	ML, data analytics, computer vision	Reduced crime and violence	Invasion of privacy and ethical implications.	◕
5.04	Cyber incident response simulation	Both	Supporting capabilities	Supporting capabilities	Use generative AI to create a dynamic cyber incident simulation that allows cybersecurity professionals to test and practice their incident response procedures.	Gen AI	Enhanced cybersecurity safeguarding sensitive data and networks	The effectiveness of the simulation is limited to the AI's training and its prompting.	◑
5.05	Summaries of security and privacy incidents	Both	Supporting capabilities	Supporting capabilities	Use generative AI to write summaries of security and privacy incidents.	Gen AI	Write summaries 51% faster while improving their quality	Potential for errors and hallucinations in summaries.	◑

Automate cybersecurity training through generative AI

AI streamlines and enhances cybersecurity preparedness and response.

INDUSTRY
K-12 Education

SOURCE
Interview (anonymous)

Challenge	Solution	Results
A large, metropolitan US school district needed to improve cybersecurity training and response exercises. The traditional methods were time-consuming, resource-intensive, and often required external consultants. The district needed a scalable solution to prepare staff for cyber threats and enhance incident response without relying on external experts.	The CISO developed a generative AI-based solution to automate cybersecurity tabletop exercises, which were traditionally run manually. Inspired by the "Backdoors and Breaches" card game, the AI created dynamic scenarios simulating cyberattacks and incident response strategies. The AI automated the creation and management of these tabletop exercises, allowing staff to engage in realistic training .	The AI-driven tabletop exercises reduced the time for need for the CISO to create the exercise (about 2 days) and the need for external consultants ($10,000 to $20,000 per session). Staff could participate in cybersecurity drills more frequently, leading to improved preparedness and response capabilities. The automated system allowed for scalable training sessions, making it possible to include more staff members without additional costs.

6 Leading AI use cases for brand impact, community engagement and social responsibility

No.	Use Case	Industry	H.E. Value Stream	K-12 Value Stream	Description	Technology	Benefits	Risks	Feasibility
6.01	AI-assisted lifelong learning assistant	Higher Ed	Teaching and learning	--	AI supports lifelong learning, helping with goal-setting, progress tracking, and feedback.	NLP, ML, Gen AI	Lifelong learning support and motivation, more effective and efficient learning	Potential bias, tracking without consent	◔
6.02	AI to support learners with disabilities	Both	Teaching and learning	Deliver Instruction	AI helps learners with disabilities access education and participate effectively.	NLP, ML, Gen AI	Increased accessibility and participation, adaptive learning experiences.	Potential bias and discrimination	◔
6.03	Chatbot for student experience	Both	Supporting capabilities	Supporting capabilities	Student questions are answered through a chatbot on the student portal relating to general front-desk capabilities.	NLP, Gen AI	24/7 support for students, reduced administrative workload	Lack of human interaction, potential bias	◑
6.04	Chatbot for mental health	Both	Student support services	Supporting capabilities	Address basic questions and mental health concerns among students and triage to human counselors for higher need issues.	NLP, ML, Gen AI	24/7 mental health support, reduced stigma	Lack of human interaction, potential bias	◔
6.05	Sentiment analysis	Both	Student support services	Supporting capabilities	A review of feedback from social media and other sources to identify positive, negative or neutral comments about the institution.	NLP, ML	Improved stakeholder insights and decision-making.	Potential biases and misinterpretation, privacy concerns	◕

Activity 1 Identify use cases that align with the institution's strategy

• Align the value drivers for education to the goals of your institutional strategy. Based on the higher-priority goals, review the use case library in this report for the corresponding value drivers.
• Download the AI Use Case Workbook for Education.
• Working with representatives from the concerned departments, identify use cases from the library that support institutional goals. These use cases will be assessed for further consideration.
• Note any other use cases of interest that you do not see in the use case library.
• Populate Tab 2, Use Cases.
  • Enter the institutional goals in column B.
  • Select the value drivers in column C that correspond to the institutional goals.
  • Select use cases in column D that you identified with your external colleagues.
• In Tab 3, Use Case Prioritization, assess each of the use cases listed in column B against the "Alignment With Strategy" criteria as either high, medium, or low.

Download the AI Use Case Workbook for Education

Input	Output
Mission, goals, objectives, and strategy Department-level goals/objectives AI use case library for education	Selected AI use cases aligned with strategy
Materials	Participants
AI Use Case Workbook for Education	Senior leadership Institutional stakeholders IT

Activity 2 Assess the use cases for risks, cost, and feasibility

1. In Tab 3 of the workbook, assess the use cases against the "Easily Mitigated Risks" criteria. This is a confidence metric. Refer to the "Risks" column of the AI use case library slides and the framework for risks in AI in education on slide nine of this report. Assess your confidence that the institution can mitigate any risks associated with the AI use case.
   • If the risk is high and you don't know how to mitigate it, choose low.
2. Next, assess the use cases against the "Affordability and Self-reliance" criteria. For each of the use cases, rank the total cost to purchase and resource the solution as low, medium, or high.
   • AI tools available as SaaS solutions are medium or high cost.
   • Those requiring custom development are low cost.
   • Those requiring ongoing service, such as a chatbot with an up-to-date knowledge base, are high cost because they require greater resources for continuous operations.
3. Finally, assess the uses cases against the "Industry Feasibility" criteria. Refer to the "Feasibility" column in the AI use case library slides.
   • If the Harvey Ball is three-quarters full, score the use case as high for feasibility.
   • If the Harvey Ball is one-quarter full, score the use case low.

Input	Output
AI use case library for education Framework for risks in AI in education in this report	AI use cases assessed for risk, cost, and feasibility.
Materials	Participants
AI Use Case Workbook for Education	Institutional stakeholders IT

Activity 3 Assess the data maturity of the relevant capabilities

1. Using the reference architecture for your respective industry, identify the capabilities that support the different use cases listed in the AI Use Case Workbook for Education.
2. Assess the data maturity of each of those selected capabilities according to the following rubric:
   • None: Data is unavailable, unreliable, duplicated, or not of sufficient detail.
   • Low: Data is available but not subject to adequate integrity or quality controls. Data ownership is undefined.
   • Medium: Low + Data is available but not fully automated. Data ownership is mostly defined.
   • High: Data is available, of high quality, and fully automated with clear ownership.
3. Heatmap the capabilities using the colors red, yellow, light green, and dark green to record your assessment.
4. In the workbook, assess each use case against the "Data Maturity of Associated Capabilities" criteria. Take an average of the heatmap assessment for the capabilities relevant to this use case.

Input	Output
AI use case library for education Knowledge of current state of data management and governance	Prioritized AI use cases
Materials	Participants
AI Use Case Workbook for Education Reference Architecture for Higher Education Reference Architecture for K-12 Education	Institutional stakeholders IT Data stewards

Leverage the Industry Reference Architectures

Use these templates to heatmap the data maturity of capabilities relevant to your AI use case.

The following templates are taken from the Industry Reference Architectures.

Download a full copy for your industry for a more complete presentation.

Download Industry Reference Architecture for Higher Education

Download K-12 Education Industry Reference Architecture

Information assessment for higher education 1/2

Information assessment for higher education 2/2

Information assessment for K-12

Activity 4 Finalize your potential list of AI use cases

The objective of this activity is to collect and evaluate a list of potential AI use cases. The outcome of this tool is a living repository consisting of a well-evaluated list of prioritized use cases to leverage with the executive leadership team/board of directors.

1. Now that you've assessed the use cases, you can review the weighting of the criteria. Based on your experience completing the assessment, do you agree that each criterion should be of equal weight? Depending on your assessment, increase and decrease the weighting of the criteria in row 7. Ensure that the total score in row 7 equals 100%.
2. The top four to six use cases with the highest percentages are the strongest candidates for further research. Conduct a market scan of these technologies to see what products may be available, which development companies have experience with the solutions, or what other institutions have implemented these tools
3. Reach out to your Info-Tech account manager, advisor, or counselor to learn how Info-Tech can further help you with this investigation.

Input	Output
Mission, goals, objectives, and strategy Department-level goals/objectives	Prioritized AI use cases
Materials	Participants
AI Use Case Workbook for Education	Senior leadership Institutional stakeholders IT

Info-Tech offers various levels of support to best suit your needs

DIY Toolkit	Guided Implementation	Workshop	Executive & Technical Counseling	Consulting
"Our team has already made this critical project a priority, and we have the time and capability, but some guidance along the way would be helpful."	"Our team knows that we need to fix a process, but we need assistance to determine where to focus. Some check-ins along the way would help keep us on track."	"We need to hit the ground running and get this project kicked off immediately. Our team has the ability to take this over once we get a framework and strategy in place."	"Our team and processes are maturing; however, to expedite the journey we'll need a seasoned practitioner to coach and validate approaches, deliverables, and opportunities."	"Our team does not have the time or the knowledge to take this project on. We need assistance through the entirety of this project."
Diagnostics and consistent frameworks are used throughout all five options.

Related Info-Tech research

Build Your Generative AI Roadmap

• A four-phased, detailed model taking you from building responsible guiding principles to executing a roadmap.
• Get a deep understanding of the generative AI landscape, risks, and opportunities.
• Review case studies and industry-specific capability maps for AI adoption.

Govern the Use of AI Responsibly With a Fit-for-Purpose Structure

• The use of AI and ML has gained momentum as organizations evaluate the potential applications of AI to enhance the customer experience, improve operational efficiencies, and automate business processes.
• Growing applications of AI have reinforced concerns about ethical, fair, and responsible use of the technology that assists or replaces human decision-making.

Identify and Select Pilot AI Use Cases

• AI has great potential but also carries great risk. How do you balance value and risk to deliver successful AI pilot projects to your organization?
• There is a sea of use cases to choose from, and everything seems important! How do you determine which use cases are worth serious time, effort, and investment?

Research contributors and experts

Pete Edwards
Enterprise Architect, Digital & Data
University of Melbourne

Pete is an outcome driven leader with over 25 years of experience in the architecture, implementation, and management of businesses, portfolios, programs, and projects in organizations of various sizes across a number of business sectors and using a variety of approaches, frameworks, and methods. His background provides a breadth and depth of capability that enables the definition and delivery of architectural strategies and roadmaps in any organizational setting.

Allister Payne
Innovation Lead,
University of Melbourne

Allister strives to create trusted and engaging digital products that help people find, understand, and self-serve with ease.

April Mardock
Chief Information Security Officer
Seattle Public Schools

April has supported cybersecurity and Info-Tech in 132 different companies. She is well versed in complex, multilayered environments, and is currently the CISO for 60,000 users. April is featured in Tribe of Hackers Blue Team: Tribal Knowledge From the Best in Defensive Cybersecurity. She holds a masters in IT and a CISSP security certification, as well as several other industry-specific certs. She is practical and pragmatic in her approaches to IT security, reducing the technobabble to manageable and actionable tasks for all users, from boardroom to lunchroom.

Dan Durkin
Managing Director of Technology
YES Prep Public Schools

Dan is a trusted strategic leader with over 25 years of experience adapting and innovating technologies to improve operational excellence. With a passion to modernize and transform current operations with next generation functionality, he is confident in balancing technical skill with business strategic initiatives to deliver value. He has industry experience in manufacturing, financial, logistics, higher education, and service-oriented organizations.

Taylor Cyr
Director
Public Sector/Higher Education at Quantiphi

Taylor is an experienced sales and program management professional with a demonstrated history of working with advanced AI, ML, and biometrics technology and an MBA in Strategy and Innovation from Boston University. He is a passionate problem-solver skilled in stakeholder and relationship management with experience leveraging clear communication skills and knowledge of advanced technology to solve cutting-edge problems in public sector and higher education.

Nicholas Burrell
Vice President,
School Partnerships at Ocelot

Colleges partner with Ocelot to efficiently answer financial aid and student service questions in ways students like – through a conversational self-service chatbot and short video explainers. Students are provided with immediate and consistent answers 24/7/365, making the topics easy to understand and reducing demands on staff.

Bibliography

"AI in Education Market Size to Reach USD 23.82 Billion by 2030 with a CAGR of 38%." Market Research Future via GlobeNewswire, 6 Apr. 2023. Accessed 27 Jul. 2023.

"AI in Education Startups in United States." Tracxn, 21 Mar. 2023. Accessed 27 Jul. 2023.

American Council of Alumni and Trustees. The Cost of Excess: Why Colleges and Universities Must Control Runaway Spending. American Council of Trustees and Alumni, 2021.

"Artificial Intelligence in Education. 2023 Survey Insights." HolonIQ, 27 Feb. 2023. Accessed 27 Jul. 2023.

Arundel, Kara. "Teachers Increasingly Embrace ChatGPT — Students Not so Much." K-12 Dive, 18 Jul. 2023. Accessed 19 July 2023.

Baytas, Claire. "How Can Universities Create AI Tools for Their Communities?" Ithaka S+R, 7 Dec. 2023. Accessed 7 June 2024.

Blake, Jessica. "American Confidence in Higher Ed Hits Historic Low." Inside Higher Ed, 11 Jul. 2023. Accessed 18 July 2023.

Brenan, Megan. "Americans' Confidence in Higher Education Down Sharply." Gallup, 11 Jul. 2023. Accessed 18 Jul. 2023.

Burke, Lilah. "Should Colleges Use AI in Admissions?" Higher Ed Dive, 18 Jul. 2023. Accessed 19 July 2023.

Butschi, Steven. "How Cloud Solutions Support Students and Schools." Google, 17 Feb. 2021. Accessed 14 July 2023.

"College Confidence Index." College Pulse, Jul. 2022. Accessed 18 July 2023.

"College Confidence Index" College Pulse via Infogram, Jul. 2023. Accessed 19 July 2023.

Cormie, Alicia. "Let's Get Personal: Adaptive Learning Tech and Education." Google, 16 Mar. 2022. Accessed 14 Jul. 2023.

"Consolidated High School District 230 Case Study." Juniper Networks, n.d. Accessed 26 Jul. 2023.

"Cost of a Data Breach 2022." IBM, 1 Aug. 2022. Accessed 18 Jul. 2023.

"Diversity, Equity, and Inclusion in Higher Education." International Business School Barcelona, 13 May 2023. Accessed 18 Jul. 2023.

"The Economic Potential of Generative AI: The next Productivity Frontier." McKinsey Global Institute, 14 Jun. 2023. Accessed 28 Jul. 2023.

Ezarik, Melissa. "7 Surprising and Significant Student Tech Use Stats to Know." Inside Higher Ed, 20 Nov. 2022. Accessed 19 Jul. 2023.

Ghimire, Aashish, et al. "Generative AI in Education: A Study of Educators' Awareness, Sentiments, and Influencing Factors." arXiv Preprint arXiv:2403.15586, 22 Mar. 2024.

Hamid, Rahem D., and Elias J. Schisgall. "Nearly Half of Surveyed Faculty Pessimistic on AI Impact in Higher Ed." The Harvard Crimson, 28 Jun. 2023. Accessed 18 July 2023.

Jackson, Nancy Mann. "Where the Tech Troubles Are." Inside Higher Ed, 31 Oct. 2022. Accessed 19 July 2023.

Kassorla, Michelle. "Teaching with GAI in Mind." EDUCAUSE Review, 14 Dec. 2023, Accessed 16 Jun. 2024.

Mardock, April. "Prompting for Cyber Incident Response Practice- A Generative AI Example." LinkedIn, 19 Jan. 2024. Accessed 16 Jun. 2024.

"Navigating Academic Life." College Pulse, Apr. 2023. Accessed 18 July 2023.

"New Poll Finds Parents Lag Behind Kids on AI and Want Rules and Reliable Information to Help Them" Common Sense Media, 10 May 2023. Accessed 27 Jul. 2023.

"NSF Announces 7 New National Artificial Intelligence Research Institutes." National Science Foundation, 4 May 2023. Accessed 27 Jul. 2023.

"NVIDIA and the University of Florida are Partnering to Bring AI to Every College." NVIDIA, n.d. Accessed 7 Jun. 2024.

O'Connell, A. J. "How (and Why) the University of Michigan Built Its Own Closed Generative AI Tools." EDUCAUSE Review, 7 Feb. 2024. Accessed 6 May 2024.

Paykamian, Brandon. "University of Georgia Advancing AI to Assess Creativity." GovTech, 17 Jul. 2023. Accessed 19 Jul. 2023.

Pelletier, Kathe, et al. 2022 EDUCAUSE Horizon Report Teaching and Learning Edition. EDUCAUSE, 2022.

"Responsible Artificial Intelligence: Recommendations to Guide the University of California's Artificial Intelligence Strategy." University of California, Oct. 2021. Accessed 19 Jul. 2023.

Rosique, Lambert, et al. "Accelerating Incident Response Using Generative AI." Google, 26 Apr. 2024. Accessed 13 June 2024.

"Teachers and Parents Report Positive Impact of ChatGPT on Teaching and Learning." Walton Family Foundation, 18 Jul. 2023. Accessed 19 Jul. 2023.

"The Time for AI in Higher Education Is Now." Quantiphi, n.d. Accessed 14 Jul. 2023.

"Virtual Agent to Support Remote Learning." Quantiphi, 2023. Accessed 26 Jul. 2023.

"With Google Cloud, Penn State World Campus to Add AI to Advising Services" Penn State University, 20 Jun. 2019. Accessed 14 Jul. 2023.