AI Policy

What you need to know

What we’ll cover:

What AI is
Examples of AI in your life
How AI works (conceptually)
ETDD’s Policy (what and why)
Examples of how to put this information into action

Goals for this Presentation

Understand at a basic level what AI is
Understand that there are different kinds of AI models
Understand different AI models are appropriate for different tasks
Be able to apply this information to using AI or writing an AI policy
Understand how ETDD’s policy addresses potential issues and be able to adapt for your organization
Have a basis for further learning

What Is AI?

How would you define AI?

“AI is whatever hasn’t been done yet.” — Tesler’s Theorem¹

Defining AI

How do you define “Intelligence”?
Academic definitions differ and are not always helpful for practical use

AI Definition Overview²
	Human-based	Ideal Rationality
Reasoning-based	Systems that think like humans	Systems that think rationally
Behavior-based	Systems that act like humans	Systems that act rationally

Defining AI

AI is “an engineered or machine-based system that can, for a given set of objectives, generate outputs such as predictions, recommendations, or decisions influencing real or virtual environments. AI systems are designed to operate with varying levels of autonomy”³

Defining AI

Questions to help determine if it’s AI:

Does the technology use data to provide predictions, recommendations, insights, or decisions?
Does the technology augment or replace human decision-making?
Does the company use words such as “personalized”, “tailored”, and “adaptive” in its marketing?
Would it be difficult or impossible to write out how a program does what it does, either because the process is opaque or because it is mathematically complex? ⁴

Examples of AI

What are some examples of AI you interact with on a regular basis?

Optical Character Recognition

Video Games⁵

Spam Filters⁶

How AI Works Conceptually

Understanding the different kinds can help you know when and how to use AI

Lets say you need an AI to help organize some recipes:

Logicist AI

Probabilistic AI

Neurocomputational AI

Think of logicist artificial intelligence as a series of rules. If you know all of your cake recipes are made with flour and butter and are cooked at 350 degrees, and none of your non cake are you can give that rule to an AI who can sort them based on the logic you gave it.

Think of probabilistic AI as showing a model a bunch of recipes labeled ‘cake’ and ‘not cake’ with notes on each ingredient; once it is trained you can show it a recipe and it’ll say how likely it is a recipe for cake or not

Think of “neurocomputational” as having a bunch of robots working in groups. You can show them a bunch of recipes, and each group will work to figure out one part of what makes a recipe. You can then ask them to label something as cake or not, or even potentially ask them to give you a new cake recipe. They could be super accurate (maybe they figured out some cakes are made with almond flour), or could be really off (they could think a recipe for wet cement is cake), because they don’t have the context to know what cake is, and aren’t good at telling you what they do / look at.

Real Life Examples of Each Kind of AI

Logicist

Probabilistic

Neurocomputational

Optical Character Recognition

Video Games⁷

Spam Filters⁸

Lots of video games use AI to control NPCs (non-playable characters). For example, in Skyrim NPCs have to know where to walk without someone controlling them, how do they do that? How do NPCs know how much they should charge for a book, or spend on a sword you want to sell them? Newer video games often use a combination of several kinds of AI, so perhaps in Skyrim there would be a logicist AI to determine where NPCs can walk, and then a neurocomputational AI to determine where they actually walk.

Unfortunately, spam emails don’t usually tell you they’re spam, so how can spam filters work? It would take hours and hours of manual work to create rules that would work for the spam I get on my work email and my personal email, much less that would be mostly accurate for most people. This is somewhere ML excels- select an appropriate model, give it labelled data, and it can create rules to catch most spam.

Optical Character Recognition is used to e.g. add words to scanned PDFs. It’s usually based on Neural nets, which are especially good at computer vision

AI Model Type Overview
	Logicist	Probabilistic	Neurocomputational
good for:	clearly defined logic	clear factors	No clear process
bad at:	handling uncertainty	operating with limited /bad data or assumptions	Truth; Explainability
example:	email inbox rules	(some) population forecasting	chatGPT

To give another example of each: In my personal life, I’ve signed up for a few store rewards cards. I like looking at the flyers before I go, but I don’t want them cluttering up my inbox, so I set inbox rules to move those to a folder. Some of the companies use different emails, so I can’t create rules that always filter correctly because there is some uncertainty about what emails should / should not be moved.

Car retailers often try to predict how many cars they will sell of certain types so they can buy appropriate stock. There aren’t clear, human definable rules for predicting that, but they have previous sales data and know they can looks at things like SUV vs truck and Honda vs Ford. The quality of data sets an upper limit on how good the forecast can be. Some of them artificially inflate demand in December by running sales, so they have to account for that outside of the Machine Learning model because the data is not accurate for December.

Computer vision uses neural nets to do things like determine if a picture is of a cat or a dog. A more practical example is to sorting old, hand-drawn paperwork. There are typically no clearly defined features to use, and no clear rules about which is which you could communicate to a computer. You can cross-train a computer vision neural network to sort the documents into different folders, but it can’t tell you e.g. why it sorted them a certain way.

Quick Clarifications

AI is not just Large Language Models⁹
AI is not new¹⁰
It is still valuable to learn new skills

AI includes more than just Large Language Models like CoPilot, it also includes things like Spell Check or the algorithms Walmart uses to know how many oranges they should order. When talking about AI (especially in the context of writing a policy), it is important to keep this in mind.

AI is not new, but it might feel that way because the history of it is full of booms and busts. We figure out something super cool, and then we over promise and underdeliver. When we haven’t solved all of the world’s problems in five years people lose interest in AI and decide all of it is useless. Then things that objectively are AI stop being called AI because otherwise they won’t get funding/interest/etc.

With the rise of new AI tools, it is still valuable to learn new skills. If you need to use complicated formulas in a spreadsheet you might be able to get CoPilot to do it for you, but if you take the time to understand the basics of what is going on you might find a better way to do it, and it’ll be easier if you have to do it again/fix it if it breaks. On the other hand, if you’re already an Excel wizard, the skill that you might find helpful is learning to get CoPilot to fix formulas when they break.

Quick Advice for Using AI

Don’t use AI search results without verifying them
If you use chat bots:
- Give them context in your prompt
- Try different prompting strategies (e.g. say “you are an X”, be nice)
Don’t forget the sunk cost fallacy¹¹
Sometimes older methods are better

AI is not a replacement for verified sources- the little robots know what a fact looks like, not what is true or not. If you e.g. look up statistics with ChatGPT, the robots will give you something that looks accurate, but you will have to double check with a reputable source (probably data.census.gov). If someone builds an AI tool that will pull census data, the source is not the tool the source will be wherever the tool gets the data from (likely data.census.gov) but you should still double check because mistakes can happen when coding things like that.

Currently, there is not clear evidence that certain strategies are universally helpful with all tasks for any given LLM model, and the models are changing at a fast enough pace that a strategy that works now might not a month from now. Instead, try different strategies to see which seems to get you what you want. Some strategies include being polite in how you word the prompt; including in your prompt something like “You are a planner,”; and prompting it to “think step by step”. This last one will use more resources, and is not always effective, so use it sparingly.

It can be easy to spend more time trying to do something an ‘easy’ way than you would doing it the ‘hard’ way. I have a tendency to do this, and what I’ve found helpful is setting a timer for a quarter or half as long as I think it’d take to do the hard way and switching to the hard way if I’m not done in that time.

AI is the new cool shiny toy that’s suddenly everywhere, and it can be exciting to use new tech (or old tech with a new coat of paint). It’s not always the most effective or easiest way to do things even if it feels like it is. There are times where AI is not the best solution. If you want to check the accessibility of a website, there is a lot you can do with automation (e.g. color contrast), but some of it is easier to just check than try to create AI for (full screen-reader compatibility).

Think about any potential questions:

Any questions you have about AI
Why an AI Policy is important
Anything specific you’re not sure how to handle

Why an AI policy is imporant

More people want to use AI
Many people don’t know how to responsibly use AI
Some AI is cutting edge technology treated as traditional software
Rapid changes in legal and technical realms
A well-written policy can save you time and headaches later

You don’t need a policy for something no one wants to use, but increasingly people want to use AI but don’t necessarily know how to properly use AI. There is a lot of conflicting and misinformation about what AI is and how to use it. Part of the issue is that AI describes a wide range of things, including some cutting-edge technology that’s being treated as if it is traditional, well-established software. One area where you see some of the potential issues with it is that the legal situation and technical capabilities are rapidly changing. A well-written policy now might save you time and issues later, whether it’s not needing to rewrite your policy a few months from now due to something like a new law or potentially saving you from legal issues due to someone unintentionally using AI in a way that is unethical or illegal.

ETDD’s Policy Outline

Introduction
Guiding Principles
Roles And Responsibilities
Definitions
General Information (and special notes)
Processes to review AI requests, sunset tools
Enforcement

Rapid Changes

Policy Considerations

Legal¹²
Technical
- Unique hacking vectors¹³
- Coding¹⁴
Consumer-Facing

The legal realm around AI is currently rapidly changing and uncertain. A lot of where we’re currently seeing these changes is with figuring out how current laws apply to new(er) AI tools. For example, multiple AI Notetakers are in court trying to determine things like if collecting biometric information without proper consent is breaking California privacy law, or if sending an AI notetaker to a meeting you don’t attend might be breaking federal wire tap law. In the technical realm generative AI is subject to unique ways of hacking traditional software is not, and cybersecurity experts are just starting to get a sense of the scope of the problem. One area of particular interest is with coding- if you use generative AI to code there are several ways it can pose cybersecurity risks that are elevated compared to coding by hand. Even if you don’t open CoPilot and ask it to code for you, if you use low or no code tools, most of those use generative AI. Coding is also a great example of how rapidly things change- since ETDD’s policy was written, the primary way of using generative AI for coding has completely changed. Even when what is legal to do or technically possible to do doesn’t change, what is available to you is probably changing. A lot of the tools you use (e.g. Microsoft Word) are probably adding in AI abilities that were not available until recently.

Rapid Changes

ETDD’s Policy

Guiding Principles
Focus on flexible processes and risk assessments
Specific as needed (e.g. AI Note takers, generating code)
Separate Generative AI Policy
Updating regularly

We include our Guiding Principles so when something changes (e.g. a new way of using AI is available, we lose access to an existing tool), we aren’t starting from square one. We already know what we care about and just have to determine how the change fits in to our existing framework. We don’t include a list of approved AI tools or use cases because it is likely things would change and then we’d have to rewrite the policy. Instead, we focus on flexible processes and risk assessments that can adapt as things change. We keep most of the policy general (e.g. all staff must be trained before using AI), but in some cases we have additional rules (e.g. AI Notetakers require consent before using, we don’t allow AI generated code) where they are needed. This lack of specificity when it isn’t needed allows us to not have to rewrite the policy as often. Generative AI is where we’re seeing a lot of these changes, so it is a seperate policy. Right now we review the AI Policy at least once per year, but can update it sooner if needed.

Lack of Accessible Information

Policy Considerations

Most information is very polarized
Most tools don’t tell you how they work¹⁵
Difficult to know if something will do what it says
Difficult to know if something has unstated risks
Lots of start ups

Most of the information out there is either very pro-AI or anti-AI, but you need a nuanced understanding of AI to write an effective policy. Most of the AI tools available don’t tell you how they work, so you don’t have a way to know if it will actually do what is says or if it has unstated risks (e.g. does that “AI coding” tool use templates or an LLM? Do you need to worry about potential malware with it?). These issues are heighted by the proliferation of start ups in the space. There are start ups run by competent, passionate people who are honest about what their product is and what it can do, but there are also start ups run by people who just care about the money who are incentivized to tell you whatever it takes for you to buy their product. Even if you find a great start up, there’s no guarantee that it’ll be here and you’ll be able to use it a year from now, and if you don’t know how it works it’ll be difficult to find a suitable alternative.

Lack of Accessible Information

ETDD’s Policy

Guiding Principles help indicate how and why
AI Training is required
AI Request process
Risk Assessments
Transparency

Guiding Principles help staff understand how they are meant to interact with AI and why. All staff must complete AI Training to ensure they have the knowledge required to properly follow the AI Policy. Staff must submit AI Requests for the AI and Technology Manager to assess- this helps make sure they are using AI in a way that aligns with our policy. Risk Assessments help ensure the potential risks are known and accounted for, and might include asking providers for additional information about models if required. These assessments can help staff understand any restrictions on their use and why. This also moves the burden of understanding AI tools and reading privacy policies etc from staff who might not understand the underlying systems or what to be concerned about to the AI and Technology Manager who can more accurately assess these tools. ETDD prioritizes AI tools that are transparent about how they work, we are transparent about how we use AI, and if we develop any AI tools we will be transparent about how they work.

Public Opinion

Policy Considerations

Public opinion divided¹⁶
Difficult to know potential consequences on relationships
Plagiarism¹⁷

Public Opinion

ETDD’s Policy

Proactively disclose our use of generative AI
Cite usage as appropriate
Avoid certain high-risk use cases
Most Risk Assessments consider this

Privacy and Data Concerns

Policy Considerations

AI Note Takers¹⁸
Data leaks
Sensitive Information leaks¹⁹
FOIA

Thinking about interacting with the public, privacy rights are a potentially major issue. Several AI Note Takers are currently in court trying to determine if they’ve violated privacy laws, and even if they haven’t a lot of people don’t want their biometric information collected without warning. Even when you’re not dealing with privacy violations, you have to worry about data leaks. Let’s say you’re in a meeting, and you’ve got a problem with one of the mayors holding up a project. And you’re honest about the problem because you need to figure out what to do. Someone in that meeting is using an AI Notetaker, and some time later that AI Notetaker has a data leak. The transcript from that meeting is now publicly available, and you could face consequences ranging from the loss of trust of your communities to potential legal consequences. And a lot of people put information into generative AI that they shouldn’t- people have put their user names and passwords, their social security numbers, all sorts of things that are now public due to data leaks. And even if you are careful about what you put in, if you give these tools access to your files, your emails, etc. it could leak anything contained in those. As a development district, anything you put into generative AI might be subject to a Freedom of Information Act Request- if this is relevant to you go ask a lawyer.

Privacy and Data Concerns

ETDD’s Policy

Don’t put information not ready for public in generative AI
Always ask for consent before using an AI Note Taker

Lack of Resources Tailored for Development Districts

Policy Considerations

Difficult to find relevant trainings
Difficult to find resources on AI policies
- Most is aimed at people who are developing AI
Time sink

Because resources for development districts don’t already exist, you will have to create ones from scratch. However you decide to write and implement an AI Policy, keep in mind it will take time. For example, when I recieve an AI request I have to spend time to figure out how someone is planning on using AI, why they’re using it, if the tool they are requesting is appropriate both for what they want and for us to use, then if it’s not I have to provide an alternative. If you want to provide blanket tools and uses you might spend time approving tools people don’t want or later having to figure out who is using it and how to handle it if there is an issue with the tool. At ETDD we approve things at the level of per person per tool per Use Case, in part to avoid wasting time researching tools for functions that haven’t been requested and to make sure people know what the policy is.

Lack of Resources Tailored for Development Districts

ETDD’s Policy

Hire a nerd
- (Who will include helpful resources at the end of this presentation)

Putting Guiding Principles into Policy

Policy Considerations

Translating Guiding Principles into policy is hard

Putting Guiding Principles into Policy

ETDD’s Policy

Example using Effectiveness:

AI is a tool not the tool
- No data analysis with generative AI
Automation Bias²⁰
Resource Use²¹

To be effective with AI, it is important to use the correct tool. This might sometimes mean using older AI tools, and often it will mean not using AI. One area in particular where some people want to use AI and shouldn’t is when it comes to using generative AI for anything factual. In ETDD’s policy we specifically state that we will not use generative AI for any form of data analysis (the tiny little robots know what a fact looks like, not what a fact is) because if you get the correct answer it will be because you got lucky.

When using AI, you have to make sure you are still outputting something that is up to the quality standards. We tend to expect computers to be smarter/more competent than we are- I think most of us have typed something in Word and gotten a spell check error, and even if you were pretty confident you spelled it correctly you probably got out your dictionary to check. It is easy to think what the computer puts out is always better than what you would create; at an extreme this can include even things like trusting information you get from generative AI more than an expert, but it can also include not being as thorough when you fact check things you get from AI because it feels more trustworthy. This is part of why you have to check any information with verified sources (also because the little robots don’t understand what a ‘fact’ is). If a model produces a bad result (e.g. bad image, inaccurate transcript, inaccurate information), you are responsible for anything that happens if you use it. AI is a tool and you are using it, it is not autonomous (yet)

While it’s less obvious when it’s on the cloud (on someone else’s computer), some AI models (generative AI in particular) use more resources than other options, and this sometimes translates to a higher cost to you. To be effective, you should balance your benefit with the increased resource use and try to find ways to reduce the increased resource use where possible.

Practical Advice for Policies

Be prepared for technical issues
Adapt for your organization
Consider how this might fit with other policies
AI is not always traditional software
Try to provide alternatives

If you have an AI incident, there is a chance it will be because someone accidentally signed up for an AI tool, or tried to disconnect it from their calendar but had a technical issue. You might be thinking that the way ETDD does things sounds like the best method, but if it involves any paperwork it will absolutely not happen at your district. If so, figure out how to adapt for your organization. You probably already have a tech policy, but your tech policy almost certainly does not cover everything you need. AI is often not just on your computer, most of the user-facing options are also online, so a policy about downloading things will have low relevance If you have to say no to something, try to understand the need to find a suitable alternative. That’ll help address genuine needs and help improve buy-in to the policy.

Putting It In Action / Hypotheticals

Let’s say you want to create an AI policy, and you just want to have the bare minimum to prevent likely issues. What do you think some of the content you would want in the policy would be? What would be the pros and cons of doing it this way?

Potential Content:

What AI can be used for

How tools are approved

How (or if) you will cite your usage of AI

How you will handle certain high-risk use cases

Other content?

Pros:

Might be easier to write and follow

Allow for staff more freedom to choose how/if to use AI

Clearer instructions

Might need to be updated less frequently

Other pros?

Cons:

Might not cover everything you need

Might need to be updated more frequently

Might not provide enough guidance

Might increase the risk of misuse of AI

Other cons?

Let’s say you wrote in your AI policy that you only had to approve each tool once, and then it was fine to use for everyone. What are some potential risks with this approach?

Approving tools for inappropriate use cases

Not approving specialized tools

Inaccurately assessing risks of a tool

Scrambling if there is a change in a tool

Other risks?

Let’s say a grant writer uses generative AI like Copilot to write a grant. How might they decide to use it?

Let’s say the grant was not funded, and the community is upset. If you had an AI policy that outlined how you used AI, how could that protect you? If you did not have an AI policy, what might you be vulnerable to?

They might use it by:

Editing their work

Asking it for statistics

Creating a sample of the grant

Creating an outline

Create the grant and not edit it

Other potential uses?

Let’s say a grant writer uses generative AI like Copilot to write a grant. How might they decide to use it?

A policy could protect you by:

Ensuring applications maintain a certain level of staff involvement

Preventing the use of statistics from unreliable sources

Helping staff use AI effectively and ethically

Other potential ways it could protect you?

Let’s say a grant writer uses generative AI like Copilot to write a grant. How might they decide to use it?

Not having a policy could make you more vulnerable to:

Using AI inappropriately

Accusations of using AI in ways that might be unethical or illegal

Relationship issues

Other potential ways it could make you vulnerable?

Resources (Policy)

GovAI Coalition²²
NIST’s AI Risk Management Framework: Tabassi Artificial Intelligence Risk Management Framework (AI RMF 1.0).
Congress.gov. “Generative Artificial Intelligence: Overview, Issues, and Considerations for Congress.” September 2, 2025. https://www.congress.gov/crs-product/IF12426.
Congress.gov. “Regulating Artificial Intelligence: U.S. and International Approaches and Considerations for Congress.” September 2, 2025. https://www.congress.gov/crs-product/R48555.
National Law Review (to see what’s in court right now): https://natlawreview.com/type-law/artificial-intelligence-ai
MIT AI Policy Briefs: in MIT Schwarzman College of Computing, n.d., https://computing.mit.edu/research/ai-policy-briefs/.
An outlook on expected AI legal issues from the National Law Review: n.d., https://natlawreview.com/article/2026-outlook-artificial-intelligence.
TN’s AI Council²³
AI legislation map: https://infographics.bclplaw.marketing/ai-legislation-map/
Article from Georgetown University on risks of AI Generated code: in Center for Security and Emerging Technology, n.d., https://cset.georgetown.edu/publication/cybersecurity-risks-of-ai-generated-code/.
OWASP’s Generative AI Security Project: “Initiatives,” in OWASP Gen AI Security Project, n.d., https://genai.owasp.org/initiatives/.

Resources (Learning more)

Introductory:

Article on AI Literacy from Stanford: n.d., https://teachingcommons.stanford.edu/teaching-guides/artificial-intelligence-teaching-guide/understanding-ai-literacy.
Stanford Encyclopedia of Philosophy’s article on Artificial Intelligence: Bringsjord and Govindarajulu “Artificial Intelligence”.
Crash Course’s course on AI: in Crash Course, n.d., https://thecrashcourse.com/topic/ai/.
Article on Large Language Models written to be accessible: in Proton, 2025, https://proton.me/blog/llm.

Advanced but Accessible:

Article on Neural Networks (this website has other articles that might be of interest): Michael Brenndoerfer Backpropagation - Training Deep Neural Networks, 2025, https://mbrenndoerfer.com/writing/history-backpropagation-deep-learning-training.
Paper on how ChatGPT works and why the responses are often generic: Ihor Rudko and Aysan Bashirpour Bonab “ChatGPT Is Incredible (at Being Average),” Ethics and Information Technology 27, no. 3 (2025): 36, https://doi.org/10.1007/s10676-025-09845-2.
Video course from MIT on AI: Rickard Brüel Gabrielsson MIT FUTURE OF AI, n.d., https://www.futureofai.mit.edu/.

In Depth:

Tensor Flow has several resources, ranging from beginner-friendly articles and courses to advanced topics with a focus on designing or using Machine Learning models: in TensorFlow, n.d., https://www.tensorflow.org/resources/learn-ml.
3 Blue 1 Brown is a youtube channel with videos explaining concepts in math and computer science: n.d., https://www.3blue1brown.com/3blue1brown.com. The videos on Linear Algebra, Neural Networks, and Probability will be of particular relevance.
Artificial Intelligence, A Modern Approach²⁴ is one of the most common textbooks on AI in undergrad

Resources (ETDD’s Policy)

Questions?

Thank you

References

2025. https://news.bloomberglaw.com/litigation/wiretap-suits-pit-old-privacy-laws-against-new-ai-technology.

n.d. https://commons.wikimedia.org/wiki/File:Email-spam-sample_(cropped).png.

n.d. https://www.microsoft.com/en-us/ai/ai-101/generative-ai-vs-other-types-of-ai.

n.d. https://owasp.org/www-project-top-10-for-large-language-model-applications/.

n.d. https://owasp.org/www-project-citizen-development-top10-security-risks/content/2022/en/CD-SEC-01-Blind-Trust.html.

n.d. https://crfm.stanford.edu/fmti/December-2025/index.html.

n.d. https://natlawreview.com/article/2026-outlook-artificial-intelligence.

n.d. https://teachingcommons.stanford.edu/teaching-guides/artificial-intelligence-teaching-guide/understanding-ai-literacy.

n.d. https://www.3blue1brown.com/3blue1brown.com.

In Center for Security and Emerging Technology. n.d. https://cset.georgetown.edu/publication/cybersecurity-risks-of-ai-generated-code/.

In Computerworld. n.d. https://www.computerworld.com/article/4041849/enterprise-note-taking-apps-face-legal-scrutiny-as-otter-hit-with-privacy-suit.html.

In Crash Course. n.d. https://thecrashcourse.com/topic/ai/.

In Google Workspace Blog. n.d. https://workspace.google.com/blog/productivity-collaboration/everyday-ai-beyond-spell-check-how-google-docs-is-smart-enough-to-correct-grammar,.

In Investopedia. n.d. https://www.investopedia.com/terms/s/sunkcost.asp.

In MIT News | Massachusetts Institute of Technology. 2025. https://news.mit.edu/2025/explained-generative-ai-environmental-impact-0117.

In MIT Schwarzman College of Computing. n.d. https://computing.mit.edu/research/ai-policy-briefs/.

In Mozilla Foundation. n.d. https://www.mozillafoundation.org/en/research/library/ai-transparency-in-practice/ai-transparency-in-practice/.

In NVIDIA Technical Blog. 2025. https://developer.nvidia.com/blog/how-code-execution-drives-key-risks-in-agentic-ai-systems/.

In Proton. 2025. https://proton.me/blog/llm.

In TensorFlow. n.d. https://www.tensorflow.org/resources/learn-ml.

Brenndoerfer, Michael. Backpropagation - Training Deep Neural Networks. 2025. https://mbrenndoerfer.com/writing/history-backpropagation-deep-learning-training.

Bringsjord, Selmer, and Naveen Sundar Govindarajulu. “Artificial Intelligence.” In The Stanford Encyclopedia of Philosophy, Spring 2025, edited by Edward N. Zalta and Uri Nodelman. Https://plato.stanford.edu/archives/spr2025/entries/artificial-intelligence/; Metaphysics Research Lab, Stanford University, 2025.

Chan, Cecilia Ka Yuk. “Students’ Perceptions of ‘AI-Giarism’: Investigating Changes in Understandings of Academic Misconduct.” Education and Information Technologies 30, no. 6 (2025): 8087–108. https://doi.org/10.1007/s10639-024-13151-7.

Chugani, Vinod. “AI Winter: Understanding the Cycles of AI Development.” In Data Camp. 2025. https://www.datacamp.com/blog/ai-winter.

Gabrielsson, Rickard Brüel. MIT FUTURE OF AI. n.d. https://www.futureofai.mit.edu/.

“Initiatives.” In OWASP Gen AI Security Project. n.d. https://genai.owasp.org/initiatives/.

Namco. English: Pac-Man Approaching a Row of Three Dots. 2024. https://commons.wikimedia.org/wiki/File:Pac-Man_eating_dots.svg.

Pasquini, Brian Kennedy, Colleen McClain, and Giancarlo. “How the u.s. Public and AI Experts View Artificial Intelligence.” In Pew Research Center. 2025. https://www.pewresearch.org/internet/2025/04/03/how-the-us-public-and-ai-experts-view-artificial-intelligence/.

Programmers at Work: Interviews Wit 19 Programmers Who Shaped the Computer Industry. Repr. with new additions. Microsoft Pr, 1989.

published, Craig Hale. “Microsoft Admits an Office Bug Exposed Confidential User Emails to Copilot.” In TechRadar. 2026. https://www.techradar.com/pro/security/microsoft-admits-an-office-bug-exposed-confidential-user-emails-to-copilot.

Romeo, Giuseppe, and Daniela Conti. “Exploring Automation Bias in Human–AI Collaboration: A Review and Implications for Explainable AI.” AI & SOCIETY 41, no. 1 (2026): 259–78. https://doi.org/10.1007/s00146-025-02422-7.

Rudko, Ihor, and Aysan Bashirpour Bonab. “ChatGPT Is Incredible (at Being Average).” Ethics and Information Technology 27, no. 3 (2025): 36. https://doi.org/10.1007/s10676-025-09845-2.

Russell, Stuart J., and Peter Norvig. Artificial Intelligence: A Modern Approach. Fourth edition, global edition. Prentice Hall Series in Artificial Intelligence. Pearson, 2022.

San José’s Information Technology Department: Digital Privacy Office, City of. AI Handbook. 2024. https://www.sanjoseca.gov/home/showpublisheddocument/109904/638463850657330000.

Tabassi, Elham. Artificial Intelligence Risk Management Framework (AI RMF 1.0). National Institute of Standards; Technology (U.S.), 2023. https://doi.org/10.6028/NIST.AI.100-1.

team, CLRN. “Is Spell Check AI?” In California Learning Resource Network. 2025. https://www.clrn.org/is-spell-check-ai/.

“Tesler’s Theorem and Other Adages and Coinages.” In Larry Tesler. n.d. https://www.nomodes.com/larry-tesler-consulting/adages-and-coinages.

Footnotes

According to Tesler, he originally said “Intelligence is whatever machines haven’t done yet”, however Tesler’s Theorem is more well known (source: “Tesler’s Theorem and Other Adages and Coinages,” in Larry Tesler, n.d., https://www.nomodes.com/larry-tesler-consulting/adages-and-coinages.)
Selmer Bringsjord and Naveen Sundar Govindarajulu “Artificial Intelligence,” in The Stanford Encyclopedia of Philosophy, Spring 2025, ed. Edward N. Zalta and Uri Nodelman (https://plato.stanford.edu/archives/spr2025/entries/artificial-intelligence/; Metaphysics Research Lab, Stanford University, 2025).
Elham Tabassi Artificial Intelligence Risk Management Framework (AI RMF 1.0) (National Institute of Standards; Technology (U.S.), 2023), NIST AI 100–1, https://doi.org/10.6028/NIST.AI.100-1.
City of San José’s Information Technology Department: Digital Privacy Office AI Handbook (2024), https://www.sanjoseca.gov/home/showpublisheddocument/109904/638463850657330000.
Programmers at Work: Interviews Wit 19 Programmers Who Shaped the Computer Industry, Repr. with new additions (Microsoft Pr, 1989)., see page 266. Image from: Namco English: Pac-Man Approaching a Row of Three Dots. (2024), https://commons.wikimedia.org/wiki/File:Pac-Man_eating_dots.svg.
Image from: n.d., https://commons.wikimedia.org/wiki/File:Email-spam-sample_(cropped).png.
Namco English.
Image from:
n.d., https://www.microsoft.com/en-us/ai/ai-101/generative-ai-vs-other-types-of-ai.
Vinod Chugani “AI Winter: Understanding the Cycles of AI Development,” in Data Camp, 2025, https://www.datacamp.com/blog/ai-winter.
in Investopedia, n.d., https://www.investopedia.com/terms/s/sunkcost.asp.
in Computerworld, n.d., https://www.computerworld.com/article/4041849/enterprise-note-taking-apps-face-legal-scrutiny-as-otter-hit-with-privacy-suit.html., 2025, https://news.bloomberglaw.com/litigation/wiretap-suits-pit-old-privacy-laws-against-new-ai-technology.
n.d., https://owasp.org/www-project-top-10-for-large-language-model-applications/.
n.d., https://owasp.org/www-project-citizen-development-top10-security-risks/content/2022/en/CD-SEC-01-Blind-Trust.html.; in NVIDIA Technical Blog, 2025, https://developer.nvidia.com/blog/how-code-execution-drives-key-risks-in-agentic-ai-systems/.
CLRN team “Is Spell Check AI?” in California Learning Resource Network, 2025, https://www.clrn.org/is-spell-check-ai/.; in Google Workspace Blog, n.d., https://workspace.google.com/blog/productivity-collaboration/everyday-ai-beyond-spell-check-how-google-docs-is-smart-enough-to-correct-grammar,.; in Mozilla Foundation, n.d., https://www.mozillafoundation.org/en/research/library/ai-transparency-in-practice/ai-transparency-in-practice/.; n.d., https://crfm.stanford.edu/fmti/December-2025/index.html.;
Brian Kennedy Pasquini Colleen McClain and Giancarlo “How the u.s. Public and AI Experts View Artificial Intelligence,” in Pew Research Center, 2025, https://www.pewresearch.org/internet/2025/04/03/how-the-us-public-and-ai-experts-view-artificial-intelligence/.
Most of the studies right now are focused on academia, but we can extrapolate from these studies that what is or is not plagiarism with regards to AI is not settled. Here is one study Cecilia Ka Yuk Chan “Students’ Perceptions of ‘AI-Giarism’: Investigating Changes in Understandings of Academic Misconduct,” Education and Information Technologies 30, no. 6 (2025): 8087–108, https://doi.org/10.1007/s10639-024-13151-7.
Craig Hale published “Microsoft Admits an Office Bug Exposed Confidential User Emails to Copilot,” in TechRadar, 2026, https://www.techradar.com/pro/security/microsoft-admits-an-office-bug-exposed-confidential-user-emails-to-copilot.
Giuseppe Romeo and Daniela Conti “Exploring Automation Bias in Human–AI Collaboration: A Review and Implications for Explainable AI,” AI & SOCIETY 41, no. 1 (2026): 259–78, https://doi.org/10.1007/s00146-025-02422-7.
in MIT News | Massachusetts Institute of Technology, 2025, https://news.mit.edu/2025/explained-generative-ai-environmental-impact-0117.
Their website is https://www.sanjoseca.gov/your-government/departments-offices/information-technology/ai-reviews-algorithm-register/govai-coalition
https://www.tn.gov/finance/ai-council.html; https://www.tn.gov/content/dam/tn/finance/aicouncil/documents/STS%20Roadmap%20for%20AI.pdf
Stuart J. Russell and Peter Norvig Artificial Intelligence: A Modern Approach, Fourth edition, global edition, Prentice Hall Series in Artificial Intelligence (Pearson, 2022).