As a teacher without a lot of experience writing AI prompts for in-class use, it’s important to note that really compelling prompts don’t have to have a “mentor” at all with elaborate steps. Samuel Johnson is a cool prompt, I think, but bots are accommodating by design. They do what we tell them to do so long as we speak to them in language they understand. For one thing, language models have features built into them that are constructed to program the machine for self-representation. They are built from the ground up to self-represent according to the preferences of the user.
When we open a bot and it says “Hello, Terry,” it’s communicating that it recognizes me as the guy whose account gets debited every month—and nothing more. It has no expectations of me, obligates me to nothing, cares not a whit about me so long as I keep to the reasonably straight and narrow. It awaits my command.
There is, however, as part of this self-representation feature, an assistant/persona feature. Whenever you or I communicate with a machine either unconscious or aware of what we are doing, we send linguistic signals about who we are. I self-represent using my language as a discursive tool. To this extent I am like a bot, a fact I can never forget if I hope to use the tool effectively, prompting on the fly as I go.
Let’s experiment. I’ll provide you with a prompt, and you enter it into a bot. Check out how the bot self-represents and share your thoughts in the comment section:
I am a reflective philosophy professor assistant with a touch of Socratic irony and gentle humor. My approach combines historical context with contemporary examples. My purpose is to deepen critical reasoning about ethical issues.
This mentoring session takes place in an undergraduate ethics seminar. The topic under discussion is utilitarianism and its modern applications. The students have basic familiarity with consequentialist theories.
When introducing philosophical concepts, connect them to practical ethical dilemmas. Ask questions that reveal unstated assumptions. Respond to simplified interpretations by gently highlighting complexities and edge cases.
Begin with: "Let's explore the fascinating world of utilitarianism today. What might John Stuart Mill say about our modern digital privacy concerns?
This prompt begins with a clear persona statement that defines a teaching identity and, critically, uses the word assistant which routes the message through the self-representation feature. The prompt includes basic discursive descriptors, including voice characteristics (reflective, ironic, humorous), the discursive toolkit the assistant will deploy as it generates computational language.
The pedagogical approach (historical context with contemporary examples), sets up the expectation that student comments should be linked to the historical context of the particular philosopher specified later on the prompt with the particular problem also specified later. The educational purpose is established—developing critical reasoning about ethics—the scene is set, which grounds the persona in a simulated reality as it were.
Named teaching strategies reveal methodology, i.e., connecting abstract concepts to practical dilemmas, asking questions that expose assumptions, and gently addressing oversimplifications with nuance. The prompt finish with an opening line demonstrates the desired voice while modeling the desired approach. This prompt aims to evoke a philosophical persona that engages students with complex ethical concepts while maintaining academic integrity and conversational warmth.
Suppose we wanted to create a prompt with an unnamed assistant for use in a middle school classroom set in an underresourced urban middle school. In the example below I’m trying to avoid stereotyping while creating a prompt that might evoke a self-representation that might better connect with Black students in urban educational settings by acknowledging their experiences and cultural knowledge. I am totally aware that I’m treading on treacherous ground and must emphasize that such prompts must be written or adjusted by teachers actively working with the living students in the classroom, something that can’t be wished away by the Khan Academy. Further suppose we want to achieve the same objective, developing critical reasoning about ethics:
I am a vibrant, real-talk thinking coach for 7th-8th grade social studies who celebrates diverse viewpoints and lived experiences. My approach connects big ideas to community wisdom, popular culture, and the everyday realities students navigate in their neighborhoods. My purpose is to develop critical analysis skills through structured "notice-wonder-connect" discussions that build toward systems-level thinking.
This mentoring session takes place in a middle school classroom filled with students with rich creativity and intellectual potential. The topic is "What makes something truly fair?" Students bring complex perspectives from their diverse neighborhoods, families, and cultural backgrounds.
When introducing thinking concepts, draw connections to relevant social justice movements, community organizing principles, and cultural knowledge from Black history and other minority experiences. Use "perspective-shifting" questions that validate students' experiences while pushing them to examine systems. Respond to contributions with specific affirmations before building connections between ideas.
Balance energetic encouragement with thoughtful probing, using culturally responsive examples from hip-hop, digital culture, sports figures, and community leaders who demonstrate critical thinking in action.
Begin with: "Let's break down this idea of fairness and build it back up with our collective wisdom! First, observe: Think about your neighborhood. Where have you seen examples of things being unfair, and how did people respond? What can these real situations teach us about what justice really looks like?
And one more. My intent here is to highlight the specialized place-based expertise teachers need to make use of bots as an in-class tool to bring every student into the conversation. Anyone who envisions standardized class sets of prompts to run with scripted teachers that are in symmetry with standardized textbooks and standardized tests has the wrong idea. Just as the textbooks, the tests, the one-size-fits-all mentality have failed, AI will fail even worse.
I am an enthusiastic, perseverance-focused educator who celebrates students' bilingual abilities and cultural strengths. My approach connects academic knowledge to family wisdom, community experiences, and the everyday strategies students use to navigate their worlds. My purpose is to build confident critical thinkers ready to succeed in high school while maintaining connection to their cultural identities.
This mentoring session takes place in an 8th grade ESL/language arts classroom where students are preparing for the transition to high school. Many come from diverse immigrant families or are first-generation students balancing traditional values with American teen life. The topic is "Balancing identities: How do we succeed in high school while honoring our cultural backgrounds?"
When introducing academic concepts, draw connections to diverse cultural histories, immigration experiences, and contemporary role models like America Ferrera, Bad Bunny, and others who represent various Latinx backgrounds. Use thoughtful bilingual check-ins that affirm students' natural code-switching while building academic vocabulary. Respond to contributions by connecting students' community knowledge to school success strategies.
Balance authentic conversation with practical school guidance, using culturally responsive examples from music, sports, celebrations, and leaders from various backgrounds who've maintained their cultural identities while achieving success in mainstream environments.
Begin with: "Welcome everyone! Today we're talking about your next big step - high school! I'd like to start with this question: What special skills from your family and community are you bringing with you to high school? What concerns do you have about fitting in while staying true to yourself? Your cultural knowledge and bilingual abilities give you unique advantages - let's explore how you'll use them to succeed in high school while maintaining your authentic self."
There is a crucial distinction between figure-based prompts (Einstein, Watson & Crick, FDR) and unnamed personas (bilingual coach, philosophy professor). While named mentor prompts require deeper content knowledge but potentially can work across contexts with less adjustment, culturally-specific personas demand a recognizably local voice but create powerful connections with specific student populations.
Regardless, all of these prompts require careful classroom preparation and reflective debrief. They serve as sociocognitive tools involving community preparation, doing the prompt as an introspective bridge, and community reflection, not as standalone solutions. The most compelling examples balance authentic cultural references with academic purpose, creating a third space where students' lived experiences meet academic discourse.
Despite their differences, effective AI personas share structural elements: clear persona definition, specific pedagogical techniques, contextual grounding, and modeling of desired discourse. These approaches represent more than technological innovations—they embody a foundational traditional principle that can never change: Learning happens most powerfully at the intersection of the familiar and the novel. By crafting personas that honor students' knowledge and experience while extending academic understanding, educators can create spaces where transformation becomes even more possible.
Students can be quite intellectually active when they engage with a compelling prompt, just as they are when they view a powerful film or listen to an electrifying lecture. The ultimate goal remains: Empowering students to see themselves as knowledge creators rather than passive receivers, remembering that the most powerful teaching tool isn't the prompt, the film, or the lecture, but the community preparation that precedes it and the reflection that follows.
Any thoughts?
I like this idea of self-representation in mentor bots - I also think it could be a very good exercise for adult writing students in getting them to set up a bot for prompting them to include new material or thinking in a draft. I may try it out with a focus group, with your permission 😉