Navigating New Frontiers: The Intersection of AI and Innovative Assessment Strategies

An Interview with Instructional Designer and Master Teacher, Nathan Shields

and

Jan 02, 2024

Have you ever had a teacher who just lit up the room with energy, insight, and creativity? Have you ever had an instructor who exuded grace and courtesy in every thought, action, and word? Have you ever had an educator who not only talked the talk but also walked the walk?

In course of my as a student, I have had the good fortune to study with several educators of this caliber: Mrs. D’Amico, my 5th-grade teacher; Mr. Siedlecki, my 11th-grade English teacher, Dr. Marcia Colish, my undergraduate Medieval Studies professor, my sitar teacher Hasu Patel, and Dr. James Phelan and Dr. Brian McHale, my dissertation advisors.

While I will resist the temptation to become too autobiographical, I can say–as many of us can–that life has had its ups and downs, but the one constant in my life–in addition to the loving support of my family–has been the presence of amazing teachers to ground me, even during those times when there felt like there was no ground beneath my feet.

In my life as an educator, I have had the good fortune to rub shoulders with many equally great teachers. If anything, when you are an educator, you realize that the profession is literally full of teachers of the caliber described above.

As students, we rarely notice this because so often we are so focused on ourselves, developing skills, and getting over the next hurdle, that we forget to take the notion of the “human capital” that serves as the essential architecture of any instructional apparatus.

When I first began Educating AI, I recorded two podcasts celebrating some great educators I more recently had the pleasure to cross paths with: Jon Graetz and Bryan Lakatos. I’d encourage you to take a listen if you haven’t already.

Today, I want to celebrate another educator who I recently started working with: Nathan Shields.

Upon first meeting, I immediately realized I was encountering someone who would not only make a profound difference at my home institution but also more broadly impact the way we think about and approach contemporary education.

Nathan has many lessons to offer the growing community of Educating AI. As curricular thought leader, he has designed a cutting-edge system for implementing and integrating standards-based assessment into a college-prep high school curriculum.

He is an early adopter of gen-AI as an powerful classroom tool and thus has pointed insights into how to best facilitate instructional conversations about AI literacy and how to develop AI-resistant prompts and assignments that push students to do work and demonstrate skills beyond the reach of current large language models.

In addition, Nathan embodies a sensible approach to contemporary that balances a rootedness in the collection and analysis of evidence and data with a deeply humanistic impulse for the social-emotional needs of post-pandemic students as they navigate a world filled with challenges very different to those most educators encountered when they were in school or even when they were receiving their teacher trainings.

In this post, I am giving Nathan the floor to write and speak in his own voice at length. Please bookmark this post. Come back to it as a resource. Share it with friends, family, and colleagues. In it, you will witness on a powerful instructional mind at work. I guarantee this is not the last time you will hear the name, Nathan Shields.

If anything, I hope that he will return to write another post here at Educating AI. If you want to relay feedback to Nathan or ask him a question, please drop a note here or send me an email at potkalitsky@gmail. I will put you in contact.

1. Introduction to Nathan Shields

Need to insert Question A

Hello! My name is Nathan Shields and I am in my eighth concurrent year as a professional, licensed educator in the State of Ohio. I was trained as a social science educator at Wright State University in Dayton, Ohio. Most of my career, at this time, is teaching the State of Ohio’s curriculum known as Modern World History at the sophomore and freshman level, but I have taught sociology in some form seven out of eight years.

The State of Ohio has no sociology curriculum recommendation, so from the start I have used a modified version of the American Sociological Association’s recommended curriculum for high school sociology. I assisted in a minor way with a college-level introduction to sociology course while I was at a local Public School district.

“The Sociological Imagination” Created by Midjourney

The sociology course as I teach it has always contained four units that are separated as such:

Unit 1: The Sociological Imagination & Inquiry Basics
Unit 2: Culture & Groups
Unit 3: Socialization & The Self
Unit 4: Stratification & Social Movements

At my previous school, I taught more theoretically about sociology, but quickly moved into having all students conduct amateur survey research on a question of their interest related strictly to the happenings with the school community. This past Fall Term of 2023, I taught sociology at my current school as a term-long inquiry into a core question of the student’s choice, with a methodology of their choice, and then implementation and data analysis that aligned with their choices.

All students then created a summary of their research and emailed it to an authentic audience of their choice. While classes can always be improved, I am happy overall with how my current more authentic, amateur research oriented sociology course was conducted and look forward to tweaking and running it again in future school years. I will be exploring more knowledge regarding how to efficiently engage students in authentic quantitative and qualitative analysis of their collected data.

As for United States history, I have now taught United States history for about five years total across my career. In my student teaching experience, I assisted in the delivery of United States History at the eighth grade level, which brings students from indigenous American through to the end of the U.S. Civil War and immediate conclusion of the Reconstruction period.

At my previous school, I was tasked with the State of Ohio’s curriculum that reviews core founding documents, but then skips up to the post-Reconstruction era.

This curriculum, as I taught it, was targeted to the sophomore level. However, due to the common issue of eighth grade educators not reaching the Civil War and Reconstruction by the end of eighth grade, I always taught a brief review of the U.S. Civil War and Reconstruction before re-aligning with the pacing expected by my district, rooted in the State of Ohio’s education standards.

As an employee of my current school, I teach a survey of all of United States history, from indigenous American to the present. Currently, this class is at the junior level, but policy change will be shifting that to the sophomore level.

In order to meet this challenge, I devised seven units of “periodization” to simplify the task of the survey course. This means that I have had to determine what key events and vocabulary to emphasize, inevitably leaving out some events and vocabulary that other United States history educators may otherwise include.

“The Biased Lenses of US History” Created by Midjourney

Unit 1: Settling the Americas (origins-1750)
Unit 2: Revolutionary America (1750-1815)
Unit 3: The Industrial Impetus (1815-1854)
Unit 4: Growing Through the Pain (1855-1900)
Unit 5: The Roars and the Reticence (1895-1932)
Unit 6: Becoming a Superpower (1933-1979)
Unit 7: The Promise of Our Nation (1980-present)

I teach my U.S. history course in a constructivist style, where I lecture very little and control the student’s investigations in minor, but important ways that keep them within the expected periodization or intentionally challenge them to create connections across time while avoiding presentism.

I find it inappropriate to dictate to students history as objective fact, because all history has been developed with bias. I strive to present them with primary sources, help them interpret them, but ultimately challenge them to construct their own interpretation of history and how it compares and contrasts with the sourcework they are relying upon.

At the sophomore level, I will retain a constructivist strategy, but plan to:

Be more restrictive in the modality the students use to present their interpretations of history
Rely more heavily on credible tertiary reading to help students grow their contextual knowledge-base
Reduce the intensity of long-term assignments to better fit the sophomore skill-set and push them into their zone of proximal development without overwhelming them.

I will unfortunately not be able to teach them the same way that I have taught the juniors, but so it goes.

Overall, I have been influenced by the College, Career, and Civic Engagement (C3) framework developed by Swan et al. and published in collaboration with the National Council for the Social Studies. I strive to give students ownership of their own historical interpretations and how to support them, prioritize development of their historical thinking skills, and engage them in how to generalize historical lessons into civic action.

1.B What inspired you to pursue a career in education?

I was inspired to pursue a career in education after receiving multiple compliments as an amateur teaching assistant for Guru Jeff Brown of the Brown Institute of Martial Arts in Centerville, Ohio.

In addition, for years I was presented with the reflections of others that I am a “great teacher” and/or “great with kids”. For these reasons, I had switched from a digital design and creative writing pathway at Sinclair Community College to a pathway of physical education.

My desire to pursue physical education took me all the way to Wright State University, within their physical education teacher preparation program, but at a certain point I realized and predicted a few things.

I did not have knowledge nor interest in obtaining much knowledge of professional or amateur sports. Physical education teachers are not as highly respected in K-12 education as their peers in other disciplines.

The training I was receiving seemed unnecessarily complex and uninteresting for a career in K-12 education, especially given that I would be working mostly with students in K-9, where physical education requirements ended.

I did not wish to pursue professional fitness training and wanted to remain in the education field. I thought back to formative experiences of my own, and decided to pursue social science education. This pathway stuck long enough for me to achieve both a bachelor’s and master’s degree from Wright State University.

2.1 How would you describe your approach to teaching?

Every student is worthy of the time, attention, and positivity required to develop a strong teacher-student relationship and a lifelong love for learning. I will do everything within my power to help a student succeed, and will treat every day as a fresh start for each and every student.

I will empower all of my students with a sense of personal responsibility and ownership over their learning. I will be open with students about exactly how they can succeed, and give them every opportunity to do so.

When difficult circumstances arise, I will maintain a positive outlook on a student’s potential for success, and present a clear pathway toward academic achievement. I will be there for students who have encountered personal struggles, and lend a listening ear and helping hand whenever I can.

I will hold students accountable. As a driven and optimistic person myself, I want to see the absolute best out of my students and myself. I will ask students to hold me accountable. I will reflect constantly so that I provide the best education possible to my students.

I will embody every great teacher who has ever taught me, so that my students will always remember that I cared, wanted and expected them to succeed, and did not take my responsibility for their learning for granted.

2.2. Could you share a specific example where your teaching philosophy significantly influenced your lesson planning or student interactions?

My decision at my previous school to provide fully completed models of long-term assignments to my students was influenced by my philosophical desire to hold students accountable for their learning, minimize excuses, push them to achieve highly, and present to them my best self as an educator.

My decision to abandon the lecture-based teaching model, especially prevalent in the instruction of history, was rooted in my desire to give students more ownership of their learning.

Students cannot own history if I am telling them what is and is not. Within the limits of good historical thinking skills, I began to develop ways for students to see themselves in the curriculum, investigate their specific interests, and still retain some general “need-to-knows” regarding historical narratives.

My experiences at my previous school encouraged me to be highly communicative with students and parents. I was driven to ensure that my planning was highly transparent and always readily available to anyone who wanted to see it.

To be on top of potential academic concerns and to ensure that I would be able to protect and defend myself from accusations of incompetence, favoritism, or other examples of parents and students deriding teachers because the student’s grades are low or their feeling of classroom engagement is low.

3. Impact of AI on Education:

3.1 When did you first learn about generative AI, and what was your initial assessment of its impact on education?

I first learned about generative AI a few years ago, perhaps around the middle of COVID-19 pandemic in full swing. I learned that AI tools were able to generate written summaries of works and could help people locate information.

Mary Reagan, Ph.D., “Understanding Fairness and Bias in AI Systems,” Mar 24, 2021

At initial assessment, these tools [Gen AI] seemed like evolutions of existing search engines. My initial assessment of AI tools for education was that it would be a more efficient, but still bias-prone method, for students to collect information, just as ordinary Internet searching is.

3.2 How has your evaluation of generative AI's impact on education changed over time?

Generative AI is a great brainstorming / searching tool. In this way, it becomes an avenue for research that can lead to higher quality sources of information.

Generative AI is a tertiary source of information, and in that category it should be treated as such within academic and professional contexts.

Generative AI is biased based upon its training. I would like to know more about generative AI localization across different languages, and how it is trained. Is it trained on translated versions of the same texts, or on texts only in the target language, or both? How are those translations produced?

Generative AI is a tool that can produce “work” of tertiary summary and potential application much as teaching asks students to do, and therefore it presents a challenge to educators to ask students to obtain and use knowledge and skills in different ways from the past. I think this challenge is a good thing for the field of education.

Generative AI can be used to shortcut the learning process in unproductive ways, creating reliance upon technology to do the thinking for us without also directing us to then commit our saved energy toward higher level learning and application. It can produce dependence if misused or overused.

3.3 In your opinion, what are some productive ways teachers can use and respond to AI in the classroom?

Teachers can use AI to:

Help students find information quickly, without having to rely upon “smart search” strategies used within ordinary search engines.
Create teachable moments for corroboration of information, pointing out generative AI’s limitations for producing accurate information.
Create outlines for assignments, questions, and activities when they are otherwise out of ideas. The generative AI toolbox and huge library of training documents becomes a more efficient way to produce educational activities and resources that are both simple and, with editing, complex.
Provide basic feedback on students’ digitally written work to shortcut the feedback process if needed. However, I would not consider this a highly effective or credible professional usage of generative AI.

Teachers can respond to AI by:

Being open-minded about how AI tools could benefit their students.
Gathering writing samples from students early and often, separate from digital typing, or as part of a “live-typing” exercise. This way there is enough background information to know if a significant leap in English writing ability should be suspicious or not.
Using it themselves in professional ways and for personal usage.
Asking for professional development on AI tools from their administrators.
Being clear and upfront with students about what assignments can, can not, should, and should not be utilizing AI generation tools, and provide authentic reasons why or why not that connect to pedagogical goals.

4. AI-Resistant Assignments and Evaluations

4.1 Have you developed any AI-resistant assignments, prompts, or evaluations? Could you give us a specific example?

Asking for students to find images, cite those images, and then explain why specific details of those images relate to a vocabulary term has proven to be, for now, an AI-resistant way of requiring students to interact with key terms and vocabulary. Essentially, I modified the Frayer model of vocabulary acquisition.

Asking students to incorporate personal details of their life experiences, within their comfort zone. In other words, making writing assignments include personalized details that an AI generated tool may do too generically. At a certain point, trying to coach the AI to do it would take longer than simply writing the sample authentically.

Attach assignments to authentic audiences or purpose such that the student wants to complete the work as a representation of their own ability.

Be clear what you are measuring with an assignment, such that the student can understand if they are or are not achieving the stated goals. This makes the student reflect on their ability rather than only on achieving the grade, and thus prioritize their own learning over shortcutting an assignment with AI in order to earn the grade.

This has not worked very well for a handful of my students who have English as a second language, and this indicates to me that I should have collected in-class writing samples (that were impossible to generate with AI) much sooner than I attempted to. I will be rectifying this issue in Spring Term 2024.

Increase the depth and complexity of an assignment, without necessarily increasing its length. In other words, keep repetitions lower and increase the quality of each repetition so that students do not feel pressured to use AI generation to save time.

Provide students with various alternative methods of delivering work products beyond the traditional essay. Integrating art, presentation skills, or complex questions that AI tools simply cannot answer in the proper context of the classroom goals.

Once I move to sophomore level U.S. history, I will likely require students to do some assignments in non-essay format, and some in essay format.

Through providing my model assignments with an account of how the work was done, students can see when, if ever, I used AI tools to help me generate an idea or an outline.

Through providing unlimited revision and resubmission of summative works until the end of term, I have encouraged students to do the work themselves rather than rely on an AI generation tool in order to save time. The double-edged sword is that students who procrastinate may still rely on AI tools when the end of the term looms.

Expect assignments to build upon one another or connect to one another over the course of a term. This makes it even harder for generative AI because, again, the student would spend almost the same amount of time refining the prompts in the AI tool when they could just do the work themselves and make the expected connections.

Here is an example of a rubric of a summative assignment in U.S. History from Fall Term 2023 that also took on the role of the end of term “final assessment”. Note the various connections to previous work and authentic outreach.

Here is an example of a summative assignment in Sociology from Fall Term 2023. Note how it builds upon the data collected in earlier units and includes a separate, personalized reflection on the research process.

Note from Nick: Check out MagicSchool’s AI-Resistant Prompt Improver for other suggestions.

4.2 How do these assignments challenge students in ways that AI cannot replicate?

I have really honed in on students being specific, and not speaking in generalities. By requiring highly specific inquiries into history driven by student interests, it makes it harder for the tertiary product of ChatGPT to earn the student much more than a D+ or C+ on a given assignment.

They then have to revise the work anyway in order to achieve higher scores. In history, this causes some trouble for students because they tend not to always get the “big picture” narratives of history, as I have them focus on sometimes pedantic details.

Mind that those “big picture” narratives are always biased and should be carefully thought through before being shared with students anyway. I am currently thinking of more engaging ways for students to absorb tertiary, contextual information and use it meaningfully in class and in their major work products.

To reiterate, the inclusion of meaningful intra and interdisciplinary connections, interaction with authentic audiences, and personalization of the learning are the three main “pillars”, if you will, of how I think about creating AI resistant assignments.

The “pillars” metaphor comes from Joe Feldman’s book, Grading for Equity, and his three “pillars” of equitable grading: (1) motivational, (2) accurate, and (3) bias-resistant.

5.1 Can you explain your outcomes-based grading system and how it differs from traditional grading methods?

I would be happy to! The outcomes-based grading (OBG/SBG) system that I use was inspired by my reading of Feldman’s Grading for Equity.

Nathan’s Stones and Buckets Analogy:

The analogy I tend to use is that traditional grading is like collecting stones to fill a one-way bucket.

The stones are assignments and the bucket represents the students overall achievement in the class.

Some stones are big while others are small. Students have to pay attention to all the stones, but sometimes let the small stones pass by because they do not matter as much at the time (homework, practice activities).

However, miss too many small stones and all of a sudden your bucket might weigh less than you need it to by the end of your collection. Forget to collect a large stone and you’re in for some real trouble (missing a test, doing poorly on a big project).

My goal is to shift conversations away from grades and points and to the content, skills, and my feedback on students’ summative work.

It is an uphill battle because students have, since third or fourth grade, been subject to the collection of stones to fill a bucket, and the more stones they collect, the better their chances of earning the coveted A+.

The quality of these stones also varies quite a bit as students encounter different teacher expectations.

I go into OBG/SBG with eyes wide open that no grading-system is perfect. All evaluation systems are a contrivance of our limited time with students and the institutional pressures to evaluate, sort, and determine each student’s relative level of knowledge, skill, and aptitude.

That said, in its current iteration, the outcomes-based grading system I use can be boiled down to four steps, and the first is the most important by far.

First page of the PDF file: Equitablegrading

Pleasanton Unified School District, “3 Pillars of Equitable Grading Practice”

Step One:
Determine the “outcomes” that will be instructed toward during the course. These statements are used to guide instructional decision making in terms of content and skills acquisition and practice for the students.

Every assignment is aligned with one or more of the outcomes that have been written well ahead of the class start date.

Example: In United States history, the Fall Term’s four units are split across 28 separate outcomes, and each unit addresses a specific number of those unique outcomes.

Then, the Spring Term’s three units are split across 24 separate outcomes, and each unit addresses a specific number of those unique outcomes. Some skill-based outcomes do duplicate across terms, but this is fine because the grading period is different.

Step Two:
Develop more specific activities that formatively assess the outcomes in each unit multiple times, informally (meaning without grade), and once, formally (meaning with a grade). Each unit includes one summative assessment where the highest score in each outcome is made available to the students.

Example:

Students take a unit pre-assessment, prepare exploration activity presentations, and then present and answer questions on a post-assessment.

Only the post-assessment is a formal score, but it will later be overwritten by any higher scores earned on the summative assessment.

The data from the post-assessment, however, drives further formative and informal targeted lessons for content and skills.

Students begin the summative assessment, which is scheduled for a typical completion time of two weeks, with 4-6 hours of dedicated student work time to the assignment overall.

Time is provided in “workshop” during class. I know the estimated work time for students because I have completed the assignments myself

Step Three:
Update student grades on a weekly basis, with written feedback on drafts of summative work. Use checkpoints to provide feedback and update grades as students display progress on summative work.

Step Four:
Determine the appropriate conversion to the traditionally reported scores of A-F. This conversion must occur every week, at midterm, at the end of a unit, and at the end of term.

The conversion must also be scalable such that it is motivational, accurate, and bias-resistant when there are only some outcomes at the start of a term, but then all outcomes cumulatively measured at the end of that same term.

The process of iteration to get this conversion right has been the most vexing yet rewarding aspect of my OBG/SBG growth, because it is like an impossible puzzle that every time I think I’ve solved every problem for, a new problem arises.

It must be relatively simple, clear, and easy to explain, but still motivational, accurate, and bias resistant.

In the next section of this reflection, I will critique my most recently used conversion system and detail the next iteration that I feel, honestly, is the best version yet. With a little luck, it may even be the final version.

Grade Conversion Systems:
This past Fall Term, I decided not to use any measure of central tendency to calculate a student’s overall achievement level.

The mean has the weakness of removing incentives to do later work if a student is satisfied with lower than maximum achievement.

It also weights outliers in the data such that achievement scores become skewed, and subjectively inaccurate.

The mode and the median both have the weakness of disincentivizing later work because students can collect enough outcomes to reach those goal numbers too early in the term.

I instead utilized what I called “Mr. Shields’ Measure of Excellence”, where the Nth number in the median set would be there grade based on the following conversion:

Nth Number
A+ = Nth Outcome at Level 5
B+ = Nth Outcome at Level 4
C+ = Nth Outcome at Level 3
D+ = Nth Outcome at Level 2
U = Nth Outcome at Level 1 / Level 0

This system created high levels of motivation, but resulted in wild swings in gradebook reports just like last year.

This system was also not subjectively or objectively accurate when students were in unique situations where their Nth-1 number was a Level 5, but their Nth number was a Level 0, at the end of the term. In the system described above, this student’s overall reported achievement would be a U (Failing)!

This was a problem with low scoring outliers skewing the grade in a manner even worse than the typical mean calculation. I was aghast and appalled at my own lack of mathematical prescience.

Thankfully, these outlier situations only happened to two or three students, and the situations were not as extreme as to lead to a failure with a plethora of other Level 5s achieved.

However, a few students who earned a C+ had so many Level 5s in their outcome outlay that a B+ was probably the more accurate score. This lack of accuracy in grade reporting was also harmful to my goal of using an increased number of high achieving outcomes as a way to differentiate embedded honors expectations.

I solicited feedback from all students at multiple points, but only two this term ever provided me with productive feedback on the conversion system.

I did some cursory searching online for some more variations of conversion systems, and ran many different, yet simple tests for how different conversion systems would respond to various outlier situations and students who were intentionally trying to “game” the system.

Thanks to these hours spent, I will be piloting a new “Measure of Excellence” in Spring Term 2024 that revises Step Four conversions with the following rules.

It is more complex than the Fall Term 2023 system, but that added complexity does yield more accuracy without harming the motivation needed to achieve an “A” and without inserting any qualitative bias into the calculation. With the system below, I truly believe I can deliver on my pocket statement to students that, “This class is easy to pass, but really hard to get an ‘A’ in.”

The Measure of Excellence
(Nth Number, Average Outcome Otherwise, and Auto U)

A = Nth outcome at Level 5 w/ no outcome below Level 3
1. A+, w/ no outcome below Level 5
2. A-, w/ any outcome below Level 3
If not meeting the “A” goal, then use the average outcome rounded down.
Conversion Chart for Average Outcome
4 = B+
3 = C+
2 = D+
U = ≥ x # of outcomes at Level 1 or Level 0
Sample Fall Term U.S. History

Example Student Progression: Units 1 and 2

Example Student Progression: Units 3 and 4

I am very excited to see how this new system results in more accurate student achievement scores as well as reducing grade inflation in the “A” range.

I love giving students grades in the “A” range, but I also want those “A” range scores to be a meaningful and difficult goal that any student can achieve given the right work ethic, engagement, and dedication to the revision and resubmission process.

5.2 How might your grading system serve as an effective strategy in response to AI integration in schools?

The OBG/SBG system of conversion I used in Fall Term 2024 was not, in my opinion, an effective strategy in response to AI integration. It created too much of a high stakes environment.

The allowance for revision and resubmission of summative work all the way to the end of term did, I feel, help students feel like they had time to commit to make the work their own.

Getting students to look at and truly consider the written feedback I leave them more consistently will be a major goal of Spring Term 2024 and the new concept of a ‘Checkpoint Reflectpoint’ that I will be trying out at the conclusion of each unit of study.

The new system that I will be piloting in Spring Term 2024 will be more of an effective strategy in response to AI integration in schools because it will reduce the “high stakes” nature of the end of the term.

Students will not experience the large and mildly unpredictable swings in their grades.

However, it will lock A-, A, and A+ behind a very specific expectation wall that I can tailor using rubric language that strategically limits the usefulness of AI generation as a replacement for authentic student work.

SBG/OBG’s emphasis on summative work helps students focus on one or two major assignments at a time, as long as they are not procrastinating. The work becomes less generalized and student responses are not expected to be the “same answer.”

Students are expected to choose various different pathways. All of this choice, as part of a constructivist pedagogy within some boundaries, helps students take ownership of their work.

Again, I have noted two or three suspicious instances of work from a handful of students this Fall Term, but the vast majority of work samples do not strike me as generated by AI thanks to their word choice, personal connections, and specificity of sourcework.

SBG/OBG’s reduction in the need to grade formative practice reduces the need to assess student work when the answers are all likely to be only slight variations on one another.

This reduces student-to-student plagiarism as well as student-to-tool plagiarism. These shared, formative understandings are valuable, of course, but not something that needs to be formally evaluated for a completion or accuracy score when less formal, more fun, and still data producing methods exist.

5.3 Could you provide an example of how this grading system has impacted student learning or engagement?

It is hard to say exactly how honest the students are when they mention how much they have learned in the class, but reports are good. I must go out of my way to be clear with students about OBG/SBG practices and why I do things the way I do.

A few students initially faced significant challenges during the first months of the Fall Term. However, instead of finding themselves in a situation they couldn't recover from, the Outcome-Based Grading/Standards-Based Grading (OBG/SBG) system I employ provided them with a clear path to achieving success, enabling them to earn an A or a B.

Were they taking advantage of a system so that they could procrastinate? Perhaps. But in the end, that matters less to me than the fact that they did the work to meet the requirements for attaining those As or Bs.

Furthermore, my even more regular checkpoint grade conversions kept low performing students at lower achievement levels, and did not artificially inflate their grades as happened in the 2022-2023 school year.

With that problem fixed in Fall Term 2023, I was able to send meaningful concern emails to students, advisors, administrators, and guardians in the lead up to major term deadlines.

I am happy to say that there were very few “surprise” grades and only two direct parental concerns that were easily addressed once they realized that their student had ample opportunity to revise and resubmit work in line with the substantive feedback I had written on Canvas earlier in the term that had simply never been acted upon by the student.

6. Reflections on Teaching Career:

6.1 Looking back on your teaching career, what has been the most rewarding experience for you?

There are two most rewarding experiences for me:

I was entrusted with creating teaching videos for all of Modern World History and United States History curriculum for high schools in my previous school during COVID-19 lockdown and remote learning. The first set of Modern World History videos I did simply because I wanted to; I needed to have something for my students and felt it was the best I could do for them at the time.

I then happily shared those videos with my district curriculum coordinator and my colleagues across the district. I was then invited to complete U.S. History videos for a stipend. I was happy to create them. I wonder if they still get used from time to time…

The second experience chronologically, but the more impactful overall, was my return to see students at my previous school’s graduation in spring of 2023 and being heartily embraced by an old student of mine. We have exchanged information from time to time, and he has kept me in the loop regarding his first semester at Wright State University.

I visited him in the Student Union to play the video game Super Smash Bros., as I ran the Super Smash Bros. club at my previous school where he was a student-member. I am proud to have evidence that I impacted that young man positively.

At that same graduation, a student whom I never had in class and whom I did not recall the name of had fond memories of me because I was a teacher who, when on breakfast duty, would stand by the door where students entered and say hello to every single one of them, every morning, unless I was assigned a different duty or had no duty.

Upon reflection, this moment reinforced within me the importance of ensuring that every single interaction I have with a student is the most positive, encouraging, kind, and meaningful interaction it can be, even if the context is one of academic or emotional distress, or disciplinary redirection.

This is a high bar for any human, and goodness knows I am not always successful, but I have evidence that this effort can leave long-lasting positive impacts on students.

6.2 How do you envision the future of education with the evolving landscape of AI and technology?

I envision the future of education within the evolving landscape of AI and technology to take on an increasingly humanistic approach.

As society finds itself struggling with connection in the age of globalization and digitization, I predict, and hope, that society will call upon the institution of education to engender human connection.

Generative AI and technology will hopefully become tools that accelerate how we help others with our time, talent, and treasure. That is my desire, and it is what I will work toward.

7. Nick’s Response

Thanks, Nathan, for sharing your many insights and this detailed overview of your AI-action plan and your outcomes-based evaluation system.

While reading your response I had three big insights:

Insight 1: There are biases in everything.

The teacher brings biases into the classroom through their own personal experiences, pedagogical training, ideological leanings, and practical choices in the moment.

In parallel, the technologies we use and engage with, and invite our students to use and engage with, create new vector fields where biases are reinforced or complicated depending upon our own instructional choices.

I really appreciated how you highlighted how the intrinsic biases that AI brings into the classroom can serve as instructional opportunities, but only within certain carefully constructed limits.

Larger questions remain regarding how teachers can continue to draw a clear line around practices like plagiarism when AI and its training has so complexly deconstructed that boundary to the extent that its reproductions are not prosecutable as patent or copyright violations.

Yet designers and artists around the world are currently expressing a deep sense of violation when they see their code or masterpieces replicated in a heartbeat by an artificial agent.

Insight Two: The Value of the Beginning and the Middle

Outcomes-based grading is an amazing tool for creating a different kind of atmosphere and a different system of accountability inside our schools and classrooms. I, for one, can attest to their powerful nature to recalibrate the focus of school away from points back to skills and competencies.

At the same time, as your critique suggests, the current emphasis in OBG–rightfully so for the sake of accuracy–on summative work tends to make it an easy target for students inclined to use AI in substitute of their reasoning and text generation.

More broadly, the emphasis on summative work tends to hollow out the beginning and middle of the instructional cycle despite our overt corrective measures.

I will be excited to hear how Nathan’s measures work to shift the overall atmosphere of the classroom. Does more accountability throughout the grading cycle bring more energy and engagement, during the beginning and middle of long work cycles?

I know that Nathan skillfully prepares his long work cycles so that students need to engage regularly and meaningfully to reach proficiency and mastery by the end.

And yet, our students still wait until the end to start these demanding projects. What this reveals to me–and surely also to Nathan–is that there are other factors at work here driving boredom, disinterest, and disengagement. But we will leave those for another time.

Insights 3: As We Change, Our Students Change

Your post reminds me that as long as teachers remain as reflective, motivated, and creative as you are, exemplary teachers will remain the most valuable resource in any educational situation.

As I contemplate the onset of 2024, which will most certainly see the rise of more advanced AI agents, included AI teaching-assistants and tutors, I cannot imagine in a million years a piece of artificial intelligence replacing the complex and beautiful human work that you do on a daily basis, just by being who you are for you students.

I really wish my readers could take a peek inside Nathan’s classroom for 45 minutes or so as I have had the good fortune to do on several occasions.

To Nathan, each student is a singular, complex, irreducible person deserving of the utmost grace and courtesy. He actually reminds me a lot of Patricia Ludick, my Montessori trainer.

She taught me to see the individual first–never the individual as an educational problem to solve or as a knowledge vacuum to fill. According to the Montessori philosophy, every child has a unique knowledge and wisdom inside them, and it is our duty as teacher to listen closely enough to hear that wisdom and help the child enunciate to the world.

When I watch Nathan work with his students, I see a similar transformation happening within him and thus within the students he works with. It is a rare and beautiful thing we teachers get to do. And if AI helps us get there, so be it. But we must not forget that the individual is always first.

Be well! Thanks for reading, Educating AI!

Nick Potkalitsky, Ph.D.

A guest post by

Nathan Shields

High School Educator

Educating AI