Touro University Embeds AI Across All Undergraduate Degrees by 2026

The Agent and the Assistant: Cultivating AI Literacy for Individuals and Institutions With TU — Ai4

Ai4

Higher education must shift from teaching basic tool mechanics to training students for bidirectional collaboration with autonomous AI agents across existing curricula.

AI-assisted summary of The Agent and the Assistant: Cultivating AI Literacy for Individuals and Institutions With TU — original source. Check the original for full context.

Source published: 2025-08-29

Detailed notes

## Background Published by the Ai4 channel on YouTube on August 29, 2025, this video features Dr. Goran Trajkovski, Director of Data Analytics at Touro University, outlining the institution's strategic vision for embedding AI literacy across higher education. According to the description, the talk explores how universities can treat AI as an interactive technology intersecting with human agency, integrating it holistically across general education, faculty training, and organizational governance.

## TL;DR Dr. Goran Trajkovski presents Touro University's comprehensive roadmap for embedding AI literacy into the undergraduate experience by shifting from an 'AI as Assistant' tool mindset to an 'AI as Agent' partnership model. He details the four core individual competencies required for students—communication, critical skepticism, ethical awareness, and interactive collaboration—and explains the three-tier AI Vanguard framework (Ignite, Surge, and Apex) designed to achieve 100% curriculum coverage by 2026. Emphasizing that 'evolution beats revolution,' Trajkovski highlights the importance of faculty development, continuous assessment, and institutional agency steered by a dedicated Provost for AI and institutional AI Council.

## Atmosphere & style The presentation has a warm, pragmatic, and collegial tone, beginning with self-effacing humor about post-pandemic stage jitters and early 1990s AI research before transitioning into an energetic, well-paced institutional briefing. Trajkovski's delivery combines visionary educational philosophy with clear, step-by-step administrative pragmatism. It feels grounded, candid, and constructive, offering an inspiring yet practical blueprint for academic transformation without reliance on hype.

## Key takeaways - [00:04](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=4s) Higher education must shift its focus from merely teaching specific AI tools to preparing students and professionals to live and work in collaborative partnership with AI. - [03:16](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=196s) The framework distinguishes between 'AI as Assistant' (tool-based execution with rigid human control) and 'AI as Agent' (relationship-based collaboration and bidirectional interaction). - [05:20](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=320s) Individual AI literacy requires training students as 'AI translators' who can clearly communicate technical capabilities, limitations, and institutional needs. - [05:20](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=320s) Critical evaluation and active skepticism are essential core competencies to detect hallucinations, biases, and flaws in persuasive generative AI outputs. - [05:20](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=320s) Ethical literacy focuses on recognizing that AI systems are not value-neutral and inherently shape power dynamics, economic equity, and societal outcomes. - [05:20](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=320s) Effective interaction skills require iterative prompting, challenging system assumptions, and preserving human agency during machine collaboration. - [08:16](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=496s) The AI Vanguard framework scales literacy across three tiers: AI Ignite (universal general education), AI Surge (major-specific disciplines), and AI Apex (senior capstone projects). - [11:17](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=677s) Touro University mapped 8 general education competencies to AI modules, targeting 100% undergraduate curriculum coverage by the end of 2026 after launching Tier 1 in Spring 2025. - [13:15](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=795s) Faculty development is critical; educators must experience AI as an agent within their own workflows before they can effectively teach students to do the same. - [13:15](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=795s) 'Evolution beats revolution' in institutional adoption—embedding AI competencies into existing, familiar courses is more effective than creating standalone AI degrees or isolated classes. - [15:18](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=918s) Institutions must cultivate organizational agency through leadership structures like an AI Council and a Provost for AI overseeing student learning, research, operations, and governance.

## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=4s) Introduction and the Urgency of AI Literacy

Dr. Goran Trajkovski opens the session by welcoming attendees to his first live, synchronous in-person event since the COVID-19 pandemic. Acknowledging a bit of stage nervousness, he jokes that he promises not to exceed his allocated 20 minutes. Introducing his background at Touro University, Dr. Trajkovski notes a personal milestone: his very first academic research paper in artificial intelligence was published back in 1995, humorously remarking that many in the room were likely not even born yet. He notes that Touro University’s headquarters are located at 3 Times Square in New York City, answering a question frequently posed by conference attendees.

Dr. Trajkovski transitions to the pressing focus of the talk: AI literacy, which he argues is no longer optional for either individuals or educational institutions. Every morning, millions of students use AI to complete homework, and millions of professionals rely on it to draft emails, analyze data, and assist in decision-making, while academic institutions scramble to understand the implications for their educational mission. The core flaw across higher education, he argues, is that leaders are asking the wrong question—focusing narrowly on how to teach people to use specific AI tools, rather than asking how to prepare people to live and work in partnership with AI.

- Speaker: Dr. Goran Trajkovski, Director of Data Analytics at Touro University (Headquarters at 3 Times Square, New York City). - Scope: Universal integration of AI literacy across 100% of Touro University's undergraduate student population. - Core philosophy: Treating AI as a fundamentally interactive technology intersecting with human behavior, agency, relationships, incentives, and organizational processes.

---

[03:16](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=196s) Core Framework: AI as Assistant vs. AI as Agent

The core conceptual shift driving Touro University’s strategy is the distinction between viewing AI as an *assistant* versus viewing AI as an *agent*. The "AI as Assistant" model represents dominant, tool-based thinking. In this mindset, users treat AI as a sophisticated calculator or word processor, focusing primarily on technical mechanics such as basic prompt engineering, platform navigation, and output formatting while keeping the human in rigid, one-directional control.

In contrast, the "AI as Agent" model represents a relationship-based, interactive paradigm. Rather than merely prompting for an output, users develop interaction literacy: understanding AI's underlying reasoning patterns, negotiating boundaries, engaging in iterative dialogue, and challenging questionable assumptions or errors. While both perspectives have utility, only the agent model prepares students for the emerging workplace where autonomous AI systems act as collaborative problem-solving partners capable of independent reasoning.

```mermaid flowchart TD accTitle: AI as Assistant vs AI as Agent Framework accDescr: Diagram comparing the tool-based assistant paradigm with the relationship-based agent paradigm.

subgraph AssistantModel["AI as Assistant (Tool-Based Thinking)"] A1["Mindset: 'How do we use this tool?'"] A2["Focus: Technical execution (prompting, UI navigation)"] A3["Relationship: One-directional control (AI as sophisticated calculator)"] A1 --> A2 --> A3 end

subgraph AgentModel["AI as Agent (Relationship-Based Thinking)"] B1["Mindset: 'How do I interact & partner with it?'"] B2["Focus: Interaction literacy (iterating, testing reasoning, pushing back)"] B3["Relationship: Shared collaboration (AI as autonomous partner)"] B1 --> B2 --> B3 end ```

---

[05:20](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=320s) Four Core Competencies of Individual AI Literacy

To translate the agent-based philosophy into measurable individual outcomes, Touro University identified four core competencies required for every undergraduate student regardless of major:

1. **Communication About AI:** The ability to clearly explain AI concepts, capabilities, limitations, and societal implications to diverse audiences. The objective is not turning students into deep technical experts, but rather training them to become "AI translators" who bridge technical possibilities with human and institutional needs. 2. **Critical Evaluation of AI Outputs:** The cultivation of active AI skepticism. Because generative AI produces fluent, highly convincing, yet potentially flawed or biased outputs, students must learn to question AI reasoning, detect hallucinations, identify biases, and rigorously cross-reference generated information. 3. **Ethical Understanding of AI's Societal Impact:** Recognizing that AI systems are not value-neutral and inherently embed biases, cultural assumptions, and power dynamics. Students analyze how AI influences power relationships, economic opportunities, and social equity to become conscientious ethical agents. 4. **Interaction Skills for Collaborating with AI:** Mastering bidirectional engagement with autonomous systems. Students learn to supply context, engage in iterative prompt dialogues, challenge system assumptions, and preserve their human agency while leveraging machine intelligence.

---

[08:16](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=496s) The Three-Tier AI Vanguard Institutional Model

Touro University structures its university-wide implementation through the **AI Vanguard Initiative**, a three-tier framework designed to scale AI literacy from foundational general education requirements through to senior-level capstone work.

```mermaid flowchart LR accTitle: Touro University AI Vanguard Three-Tier Framework accDescr: Flowchart illustrating the progression through Tier 1 Ignite, Tier 2 Surge, and Tier 3 Apex.

T1["Tier 1: AI Ignite Universal General Education (Writing, Math, Ethics, Jewish Studies)"] --> T2["Tier 2: AI Surge Discipline-Specific Major Courses (Business, Healthcare, Computer Science)"] T2 --> T3["Tier 3: AI Apex Senior Capstone Projects (Real-world problem solving with AI)"] ```

- **Tier 1: AI Ignite (Universal Foundation):** Embedded directly across general education courses that all students are required to take. Rather than creating new standalone classes, existing courses (such as College Writing, College Math, Fundamentals of Computers, Ethics, and Jewish Studies) are modified to incorporate core AI competencies. - **Tier 2: AI Surge (Disciplinary Applications):** Tailors AI integration to specific academic fields. For instance, business majors study AI-driven marketing analytics and customer behavior modeling; healthcare students examine clinical diagnostics and patient care workflows; computer science students delve into machine learning architecture and deployment. - **Tier 3: AI Apex (Capstone Integration):** Mandates that senior capstone projects incorporate AI tools or directly address domain-specific AI challenges, ensuring students demonstrate real-world mastery before graduating.

---

[11:17](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=677s) Implementation Milestones, Timeline, and Metrics

Dr. Trajkovski reviews the operational timeline and institutional metrics driving Touro's rollout. Touro mapped its general education curriculum's 8 core institutional competencies directly into corresponding AI-focused modules. The active rollout commenced in spring 2025 with Tier 1 (AI Ignite). By the subsequent terms, integration moves into major-specific courses (AI Surge) and capstones (AI Apex), targeting 100% undergraduate curriculum coverage by the end of 2026.

Dr. Trajkovski emphasizes that the most significant metric is not on the slides: the hundreds of hours invested by faculty members who have stepped forward to rethink pedagogical methods, redesign assessment instruments, and personally learn how to partner with AI.

---

[13:15](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=795s) Key Insights for Institutional AI Adoption

Dr. Trajkovski shares three practical insights gained from Touro’s implementation that can guide other institutions:

1. **Faculty Development is Critical:** Instructors cannot teach AI literacy if they lack robust institutional support. Touro established an **AI Speakers Bureau** along with structured faculty mentorship and training programs. Instructors must experience AI as an agent in their own workflows before they can effectively teach students to do so. 2. **Evolution Beats Revolution (Gradual Integration):** Rather than creating separate, standalone AI degrees or isolated new courses, Touro embedded AI concepts into existing courses that faculty were already comfortable teaching. "Evolution beats revolution in higher education," Dr. Trajkovski notes. 3. **Assessment and Continuous Improvement:** AI literacy represents a moving target as technology rapidly advances. Continuous feedback loops between students and faculty are essential to adjust curricula dynamically and maintain program sustainability.

---

[15:18](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=918s) Institutional Agency, AI Council, and Concluding Vision

Extending the agent framework to organizational governance, Dr. Trajkovski explains that institutions must act as "institutional agents" rather than merely deploying AI as an administrative assistant for simple process efficiencies. To steer this organizational transformation, Touro University appointed a Provost for Artificial Intelligence and established an institutional AI Council.

```mermaid flowchart TD accTitle: Touro University AI Governance Pillars accDescr: The four focus domains managed by Touro University's AI Council.

Council["Touro University AI Council Led by Provost for AI (ai.touro.edu)"]

Council --> D1["1. Student Learning"] Council --> D2["2. Research Innovation"] Council --> D3["3. Organizational Transformation"] Council --> D4["4. Institutional Governance"] ```

Dr. Trajkovski invites attendees to explore their initiatives online at `ai.touro.edu` and emphasizes that Touro's framework is fully transferable across the educational spectrum—from K-12 districts to community colleges and major research universities. Keeping his promise to conclude within the 20-minute limit, he invites interested attendees to meet him outside the hall for deeper discussions regarding implementation mechanics, assessment strategies, and faculty development.

## Notable quotes > [00:04](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=4s) AI literacy is no longer optional—not for individuals and not for institutions. — Dr. Goran Trajkovski

> [00:04](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=4s) AI is not just another (albeit powerful) tool; it is a fundamentally interactive technology that intersects with human behavior, incentives, and relationships. — Dr. Goran Trajkovski

> [00:04](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=4s) The core flaw across higher education, he argues, is that leaders are asking the wrong question—focusing narrowly on how to teach people to use specific AI tools, rather than asking how to prepare people to live and work in partnership with AI. — Dr. Goran Trajkovski

> [03:16](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=196s) Rather than merely prompting for an output, users develop interaction literacy: understanding AI's underlying reasoning patterns, negotiating boundaries, engaging in iterative dialogue, and challenging questionable assumptions or errors. — Dr. Goran Trajkovski

> [05:20](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=320s) The objective is not turning students into deep technical experts, but rather training them to become 'AI translators' who bridge technical possibilities with human and institutional needs. — Dr. Goran Trajkovski

> [05:20](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=320s) Recognizing that AI systems are not value-neutral and inherently embed biases, cultural assumptions, and power dynamics. — Dr. Goran Trajkovski

> [11:17](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=677s) The most significant metric is not on the slides: the hundreds of hours invested by faculty members who have stepped forward to rethink pedagogical methods, redesign assessment instruments, and personally learn how to partner with AI. — Dr. Goran Trajkovski

> [13:15](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=795s) Instructors must experience AI as an agent in their own workflows before they can effectively teach students to do so. — Dr. Goran Trajkovski

> [13:15](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=795s) Evolution beats revolution in higher education. — Dr. Goran Trajkovski

> [15:18](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=918s) Institutions must act as 'institutional agents' rather than merely deploying AI as an administrative assistant for simple process efficiencies. — Dr. Goran Trajkovski

## Mentioned Dr. Goran Trajkovski, Touro University, Ai4, AI Vanguard Initiative, AI Ignite, AI Surge, AI Apex, Touro University AI Council, AI Speakers Bureau, 3 Times Square, ai.touro.edu

## Video description Goran Trajkovski, Director, Data Analytics, Touro University

Abstracts

AI literacy is no longer optional—not for individuals and not for institutions. In this talk, I will discuss Touro University’s vision for weaving AI literacy throughout the university, and in doing so, reinvigorate higher education. The key is that AI is not just another (albeit powerful) tool; it is a fundamentally interactive technology that intersects with human behavior, incentives, and relationships. As we work with and adopt AI, our eyes must remain on human agency, relationships, and processes—we cannot think about technology alone. AI literacy is not just about understanding how to use AI, but also how people interact with it and how to effectively integrate it holistically into institutions. I will outline our approach to understanding AI literacy through this lens, and how we are implementing it at Touro – in our general education curriculum, faculty and staff training, and an AI-aware technology acquisition process. Our goal is to ensure that our graduates, faculty, staff, and university as a whole all become effective and critical users of this powerful new technology and are thus positioned to flourish in the emerging world of ever-accelerating innovation.

[Source](https://www.youtube.com/watch?v=bzzGDS3jpjs)

中文简报

Touro University将于2026年前将AI全面融入所有本科学位课程高等教育必须从教授基础工具操作，转向在现有课程中培养学生与自主AI智能体开展双向协作的能力。 Goran Trajkovski 博士介绍了 Touro University 将AI素养全面融入本科教育体验的系统性路线图，推动教学理念从“AI作为助手”的工具思维转向“AI作为智能体”的伙伴模式。他详细阐述了学生必备的四大核心个人能力——沟通、批判性质疑、伦理意识与交互协作，并解析了旨在到2026年实现100%课程覆盖的三级“AI先锋”（AI Vanguard）框架（启动 Ignite、激增 Surge 与登顶 Apex）。Trajkovski 强调“渐进演变优于剧烈变革”，同时指出了师资发展、持续评估以及由专职AI教务长和校级AI委员会统筹的机构能动性的重要性。 Touro University 提出了从“AI即助手”（单向的工具型使用）向“AI即智能体（Agent）”（双向的关系型协作）的范式转变，强调高等教育应注重培养人机协同伙伴关系，而非仅局限于特定工具的操作教学。 (3:16) 个人 AI 素养确立了四大核心能力：担任技术与需求桥梁的“AI 翻译官”沟通能力、主动辨识幻觉与偏见的批判性评估能力、对社会公正与权力动态的伦理认知，以及保持人类主导权的交互技能。 (5:20) 该校推出了三阶“AI Vanguard”全校推广架构，自底向上划分为通识基础层（AI Ignite）、专业定制层（AI Surge）以及毕业设计顶峰层（AI Apex）。 (8:16) Touro 将通识教育的 8 项核心能力与 AI 模块精准对齐，已于 2025 年春季启动第一阶段，并计划在 2026 年底前实现本科课程 100% 的 AI 素养全面覆盖。 (11:17)

Reading summary

Dr. Goran Trajkovski of Touro University argues that higher education must shift from teaching artificial intelligence as a passive technical tool to preparing students and organizations for collaborative partnership with autonomous AI systems. Shifting From AI Assistant to AI Agent (3:16) Higher education's predominant approach to artificial intelligence relies on an "AI as Assistant" model. In this framework, users view AI as a sophisticated calculator or word processor, focusing narrowly on mechanical tasks such as basic prompt engineering, platform navigation, and output formatting while maintaining rigid, one-directional control. Touro University's curricular strategy centers instead on an "AI as Agent" paradigm. This relationship-based model emphasizes interaction literacy, training students to treat AI as an autonomous partner. Under this approach, users learn to understand machine reasoning patterns, negotiate operational boundaries, engage in iterative dialogues, and actively challenge questionable system outputs and assumptions. Four Foundational Competencies for Undergraduates (5:20) To implement the agent philosophy, Touro established four universal competencies required across all undergraduate majors. The first, communication about AI, trains students to act as translators who can explain technical capabilities, limitations, and societal implications to non-technical stakeholders without requiring deep programming expertise. The remaining three competencies prioritize critical judgment and active collaboration. Students cultivate critical evaluation skills to identify hallucinations, bias, and flawed logic in fluent AI outputs. Ethical understanding addresses the non-neutral nature of algorithmic systems, examining how AI impacts power dynamics, economic opportunity, and social equity. Finally, interaction skills teach students to engage in bidirectional dialogue, supply context, challenge premises, and maintain human agency during collaboration. The Three-Tier Vanguard Rollout Framework (8:16) Touro operationalizes these competencies through its AI Vanguard Initiative, a scaffolded three-tier structure spanning the undergraduate lifecycle. Tier 1, designated AI Ignite, embeds AI modules directly into required general education courses—including College Writing, College Math, Fundamentals of Computers, Ethics, and Jewish Studies—mapped against eight core institutional competencies. This foundational tier launched in spring 2025. Subsequent tiers advance into disciplinary specialization and senior synthesis. Tier 2, AI Surge, integrates targeted applications into academic majors, such as marketing analytics in business, diagnostic workflows in healthcare, and system architecture in computer science. Tier 3, AI Apex, mandates the application of AI tools or the investigation of domain-specific AI challenges in senior capstone projects. The institution targets 100% undergraduate curriculum coverage by the end of 2026. Faculty Enablement and Institutional Governance (13:15) Touro's implementation relies on an evolutionary rather than revolutionary strategy, embedding AI concepts into existing courses that faculty already teach rather than creating standalone degree programs. Because instructors cannot effectively teach AI interaction without firsthand experience, the university introduced an AI Speakers Bureau, mentorship programs, and hundreds of professional development hours focused on pedagogical and assessment redesign. At the organizational level, Trajkovski argues that institutions must act as agents rather than using AI merely for administrative efficiency. Touro established an institutional AI Council led by a dedicated Provost for Artificial Intelligence, organizing its strategy around four operational pillars: student learning, research innovation, organizational transformation, and institutional governance. Trajkovski presents Touro's tiered model as transferable across the educational sector, from K-12 systems to major research universities. However, he notes that maintaining effective AI literacy remains an ongoing challenge, requiring continuous feedback loops between faculty and students to adjust curricula dynamically as underlying technologies evolve.

精炼阅读摘要

Touro University数据分析总监Goran Trajkovski博士主张，当前高等教育的核心缺陷在于将AI教学窄化为具体工具的操作训练；高校亟须从单向控制的“AI作为助手”转向双向互动的“AI作为智能体（Agent）”范式，并将全面的AI素养系统化融入全校100%的本科生培养体系中。范式重塑：从工具助手到自主智能体 (3:16) Trajkovski博士分析指出，高等教育领域目前普遍受制于将AI视作“助手”（Assistant）的工具导向思维。在这种认知模式下，使用者通常仅把AI视为高级计算器或文字处理工具，学习重点高度局限于提示词格式调整、平台操作导航等表面技术执行，人机关系处于严格且单向的机械控制状态。 Touro University提出的核心战略则是转向“AI作为智能体”（Agent）的互动关系范式。该范式聚焦于深度交互素养（Interaction Literacy），要求使用者不仅能获取生成结果，更要掌握AI底层的推理模式，学会在迭代对话中协商任务边界，并主动检验和反驳系统的潜在错误假设。Trajkovski博士强调，尽管工具型助手在日常任务中具有实用性，但唯有智能体范式才能帮助学生做好充分准备，在未来职场中与具备自主推理能力的AI系统建立平等且深入的问题解决伙伴关系。本科生个人AI素养的四大核心能力维度 (5:20) 为将智能体协作哲学落实为明确的教学产出，Touro University制定了面向全校所有本科专业学生的四大核心AI素养能力。首先是“AI沟通能力”（Communication About AI），该能力旨在将学生培养为连接技术可能与组织需求的“AI转译者”，能够向不同专业背景的公众清晰阐释AI的概念、能力边界与社会影响，而非将其局限为深度技术工程人员。其次是“对AI输出的批判性评估”（Critical Evaluation of AI Outputs），即建立主动的怀疑机制，训练学生识别模型幻觉、辨析算法偏见并严格交叉核验信息来源；第三是“对AI社会影响的伦理理解”（Ethical Understanding of AI's Societal Impact），引导学生深刻理解AI系统并非价值中立，其内部深植文化假设与权力结构；第四是“与AI协同的交互技能”（Interaction Skills for Collaborating with AI），重点强调通过多轮迭代对话与情境补充开展双向互动，在充分借助机器智能解决复杂问题的同时，始终保有核心的人类自主权。 AI Vanguard三层递进式课程整合体系 (8:16) 在教学落地架构上，Touro University构建了名为“AI Vanguard”的三层递进实施模型，将AI素养教育自通识基础纵向贯通至毕业设计阶段。该架构将大学既有的8项通识教育核心素养与AI专项模块进行精准映射，第一层为通用基础层“AI Ignite”，不设立孤立的新课程，而是直接整合进全体本科生必修的通识课程中，包括大学写作、大学数学、计算机基础、伦理学以及犹太研究等既有科目。第二层为专业深化层“AI Surge”，依据学科特征定制差异化的AI应用教学，例如商科专业着重学习AI驱动的营销分析与客户行为建模，医疗健康专业探讨临床诊断与护理工作流，计算机科学专业则深究机器学习架构与工程部署。第三层为毕业顶峰层“AI Apex”，硬性要求学生的毕业设计（Capstone Projects）必须整合AI工具或直接解决领域前沿的实际AI挑战。校方规划于2025年春季全面启动第一层通识课程改革，并计划在2026年底前达成全校100%本科生课程覆盖。渐进式变革路径与机构治理支撑框架 (13:15) 总结其实践经验，Trajkovski博士提出了推动高校AI转型的关键策略，即“在高等教育中渐进演进优于激进革命”（Evolution Beats Revolution）。Touro University选择将AI理念融入教师已有成熟教学经验的课程中，而非贸然创设独立的AI专业或学位。与此相配套，学校将师资培训置于首要地位，通过组建AI演讲局（AI Speakers Bureau）、推行结构化师徒带教体系，并投入数百小时的教师专业研修，确保教职人员自身先在日常工作流中掌握AI协同，进而具备有效的教学指导能力。在顶层治理方面，Trajkovski博士主张高校自身必须作为“机构智能体”（Institutional Agent）统筹运作，不仅局限于将AI用于辅助行政流程的简单提效。为此，Touro University正式任命了人工智能教务长（Provost for Artificial Intelligence），并成立覆盖全校的AI委员会（AI Council），全面协同管理学生学习、科研创新、组织变革和机构治理四大支柱维度的系统化转型。 Trajkovski博士认为，Touro University的AI素养推进机制与治理范式具有高度通用性，能够灵活迁移至从K-12学区、社区学院到大型研究型大学等各类教育机构；但他也同时明确指出，AI素养是一项目标处于动态演进之中的长期系统工程，高校必须在师资与学生之间构建敏捷的持续反馈机制，以紧跟技术前沿迭代课程与考核标准，保障机构教育使命的长期可持续发展。

Detailed notes — zh

## Background 本视频由 Ai4 于 2025 年 8 月 29 日发布在 YouTube 平台上，内容为图罗大学（Touro University）数据分析总监 Goran Trajkovski 博士的主题演讲。根据页面简介，演讲探讨了图罗大学如何将 AI 素养系统性地融入全校通识课程、教职员工培训以及技术采购流程，倡导将 AI 视为一种与人类行为深度交织的交互性技术，从而全面赋能并重塑高等教育生态。 ## TL;DR Goran Trajkovski 博士在演讲中阐述了“AI 助手”与“AI 智能体”的范式差异，指出高等教育应从单纯教授“工具使用”转向培养与 AI 协同共存的“交互素养”。图罗大学推出了涵盖通识教育“启燃（Ignite）”、专业深度应用“激增（Surge）”以及毕业设计实战“巅峰（Apex）”的“AI Vanguard”三层推进架构，计划于 2026 年底前实现本科课程全覆盖。此外，他强调教师赋能是组织变革的核心基石，高校应采取渐进融合策略，并在机构层面围绕学生学习、科研、组织变革与治理四大支柱构建可持续发展的 AI 伙伴生态。 ## Atmosphere & style 整场演讲氛围沉稳务实且极具前瞻启发性，演讲者以幽默自谦的开场迅速拉近与观众的距离，随后层层剖析高等教育面临的时代命题。叙事风格严谨清晰，既有对人机关系深层次哲理的精准思辨，又兼备图罗大学一手落地实践的详实细节与路线图。聆听过程给人一种摆脱技术焦虑后的从容感，充分展现了教育管理者在拥抱前沿科技变革时的笃定思考与温度。 ## Key takeaways - [00:04](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=4s) AI 素养对个人和教育机构而言已不再是可选技能，而是立足未来的核心根本，必须实现全员覆盖。 - [01:26](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=86s) 教育机构的核心问题应从“教大家使用 AI 工具”转变为“如何让人们为与 AI 共同生活和合作做好准备”，焦点需放在人的能动性、人际关系与工作流程上。 - [03:16](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=196s) 提出 AI 作为“助手（Assistant）”与作为“智能体（Agent）”的对比框架：前者停留在工具操作，后者注重关系构建、双向互动与协作共创。 - [05:20](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=320s) 确立本科生四大核心 AI 能力素养：AI 沟通表达能力、AI 输出的批判性评估、AI 社会影响的伦理理解，以及与 AI 协作的交互技能。 - [08:03](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=483s) 构建三层渐进式“AI Vanguard”推进架构：从通识教育全员覆盖（AI Ignite），到各专业深入结合（AI Surge），再到毕业设计综合实战（AI Apex）。 - [12:08](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=728s) 图罗大学计划在 2026 年底前完成全校 100% 现有本科课程的升级改造，使其全面契合 AI 能力标准，且不给师生增加额外学分负担。 - [13:15](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=795s) 教师发展是全校 AI 转型最关键的基石；教师自身必须先亲身体验将 AI 作为智能体协作的过程，才能有效引导学生。 - [13:15](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=795s) 在高等教育课程整合中“演进优于革命”，将 AI 概念融入已有既有课程比盲目新建纯 AI 课程更有效且可持续。 - [15:18](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=918s) 高校推进 AI 本质是一场深刻的组织变革，高校应以“机构智能体”的姿态拥抱变革，重新想象教育未来，而非仅充当“机构助手”修补流程。 - [15:18](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=918s) 设立全校 AI 转型四大支柱：学生学习、学术科研、组织变革与制度治理，并专门设立主管 AI 的教务长及全校委员会。 - [16:30](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=990s) 无论是 K-12 还是大型研究型大学，核心挑战均在于培养整个学术共同体在人机协同中实现共生共荣的能力。 ## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=4s) 引言与个人背景：AI 素养的紧迫性演讲者 Goran Trajkovski 博士代表图罗大学（Touro University）开场，并分享道这是他在新冠疫情后参加的首场线下同步活动，现场心情略显紧张，但他承诺严格遵守 20 分钟的发言时间。Trajkovski 博士提到自己的学术背景时透露，早在 1995 年他就发表了第一篇关于人工智能的研究论文，调侃现场可能很多人那时候都还没出生。 ```mermaid flowchart TD accTitle: AI Vanguard 体系三层进阶架构 subgraph S3["第三层：AI Apex 巅峰"] N3["毕业设计/顶点项目实操"] end subgraph S2["第二层：AI Surge 激增"] N2["各专业领域深度应用"] end subgraph S1["第一层：AI Ignite 启燃"] N1["通识教育全员覆盖"] end N1 --> N2 N2 --> N3 ``` 他强调，AI 素养（AI Literacy）已经成为所有教育机构刻不容缓的课题。图罗大学目前正在推行一项全校性计划，旨在将 AI 无缝融入本科课程体系，实现对全校 100% 本科生的全覆盖。 - **演讲者身份**：Goran Trajkovski 博士，图罗大学数据分析总监（总部位于纽约市时代广场 3 号）。 - **学术资历**：首篇 AI 研究论文发表于 1995 年。 - **核心主旨**：AI 素养不再是可选的附加技能，而是个人和机构的立足根本。 --- ### [01:26](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=86s) 视角转换：从“工具使用”到“人机伙伴关系” Trajkovski 博士直言，每天早晨全球有数以百万计的学生使用 AI 完成作业，专业人士使用 AI 撰写邮件、分析数据甚至辅助决策，而绝大多数教育机构仍在手忙脚乱地探索这对教育使命究竟意味着什么。然而，许多机构在一开始就问错了问题——他们反复思考“如何教大家使用 AI 工具”，而真正核心的问题应该是“如何让人们为与 AI 共同生活和合作做好准备”。这种视角的转变彻底改变了教学的内容、方式以及教育的意义。AI 不是普通的工具，而是一种从根本上具备交互性的技术，与人类行为、激励机制和人际关系深度交织。在拥抱和整合 AI 时，教育的核心必须始终锚定在人的能动性（Human Agency）、人际关系和工作流程上。 --- ### [03:16](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=196s) 核心框架：AI 作为“助手”与 AI 作为“智能体”的对比 Trajkovski 博士阐述了重塑图罗大学教育理念的核心框架：将 AI 视为“助手”（Assistant）与将 AI 视为“智能体”（Agent）之间的本质差异。 | 维度 | AI 助手模型（Assistant Model） | AI 智能体模型（Agent Model） | |:--- |:--- |:--- | | **思维范式** | 基于工具的机械思维（Tool-based） | 基于关系的交互思维（Relationship-based） | | **核心问题** | “我们该如何使用这个工具？” | “我们该如何与它进行协同交互？” | | **技能重心** | 提示词工程、平台操作、输出格式调整 | 交互素养、理解推理模式、协商界限、协同合作 | | **人机定位** | 人类完全主导，AI 充当复杂计算器或文本处理器 | 人与 AI 结成伙伴，共同推进任务流程 | | **典型场景** | 教学生写出更好的提示词以获取标准答案 | 与 AI 对话、质疑其错误假定、迭代想法、辨别置信度 | 尽管掌握工具层面的技能依然有用，但唯有培养智能体层面的交互素养，才能真正让学生适应未来充满自主性、创造力和独立推理能力的 AI 工作环境。 --- ### [05:20](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=320s) 本科生必备的四大核心 AI 素养能力图罗大学为所有本科生确立了四项跨专业的核心能力素养，并将其直接嵌入通识课、专业课和毕业设计中： 1. **AI 沟通表达能力（Communication about AI）**：能够向不同受众清晰阐明 AI 概念、工作机制、能力边界及潜在影响，成为连接技术可能性与人类需求之间的“AI 翻译官”（AI Translators）。 2. **AI 输出的批判性评估（Critical Evaluation of AI Outputs）**：面对 AI 生成的极具说服力但可能错误的内容，培养鉴别推理缺陷、发现潜在偏见和交叉核对信息的能力，成为理性的“AI 怀疑者”，区分 AI 的真知灼见与看似自信的谬误。 3. **AI 社会影响的伦理理解（Ethical Understanding of AI Societal Impact）**：理解 AI 并非价值中立，它内嵌了价值观、假设和偏见，深刻影响权力格局、经济机会与社会互动，引导学生成为有道德担当的参与者。 4. **与 AI 协作的交互技能（Interaction Skills for Collaborating with AI）**：学会提供上下文、展开富有成效的双向对话、共同迭代方案，在借力 AI 强大能力的同时牢固保持人类自身的主体性。 --- ### [08:03](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=483s) 图罗大学“AI Vanguard”三层推进架构与落地时间表图罗大学的“AI Vanguard”（AI 先锋）先导计划采用三层渐进式架构，将 AI 素养从全员普适基础层层推进至专业深度应用： - **第一层：AI Ignite（启燃）**——嵌入通识教育课程。所有本科生无需额外选课，直接在现有的必修课程（如大学数学、大学写作、计算机基础、伦理学基础、犹太研究等）中融入 AI 概念，实现通识教育与 AI 素养的融合。 - **第二层：AI Surge（激增）**——深入各专业领域。商科学生探索营销分析与消费者行为预测，医疗健康类学生研究医学诊断与护理优化，计算机专业则深钻机器学习底层实现，拒绝“一刀切”式教学。 - **第三层：AI Apex（巅峰）**——毕业设计（Capstone Project）实战应用。所有大四毕业设计必须结合 AI 工具或解决所在领域的 AI 相关挑战，检验学生在跨学科复杂情境中将 AI 作为问题解决伙伴的综合能力。 Trajkovski 博士表示，该计划的优势在于充分依托现有教学架构，不给师生增加多余的学分负担。学校从春季学期开始重点推进通识教育层的 AI Ignite 模块改造；下一学期将推行各专业的 AI Surge 模块；随后进入毕业设计阶段的 AI Apex，预计在 2026 年底前实现全校本科课程体系的全方位覆盖。在具体模块开发上，团队对照通识教育原有的 8 项核心能力进行了深度拆解与 AI 场景化转换。 ### [12:08](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=728s) 推进时间表与教师发展在 AI 变革中的核心作用图罗大学（Touro University）在 AI 领域的相关举措正式启动于 2025 年。演讲者坦言，学校和所有人一样，正处于摸索前进道路的起步阶段。根据规划，学校预计到 2026 年实现 100% 的课程覆盖率，完成现有课程的升级与改造（uplifting），使其全面与相关的 AI 能力标准保持一致。然而，幻灯片上最关键的数据往往不是冷冰冰的百分比，而是教师们在幕后所付出的心血。教职员工们已经投入了数百个小时来重新思考教学模式、开发全新的评估考核机制，并学习如何由自身去与 AI 建立协作伙伴关系。演讲者特别强调，教师发展（Faculty Development）是支撑全校整个 AI 转型进程中“绝对至关重要的核心”。 ```mermaid flowchart LR accTitle: 图罗大学 AI 落地推进路线图 subgraph Phase["推进路线图"] A["2025 年启动规划 摸索实践路径"] --> B["教师发展与课程重构 投入数百小时"] B --> C["2026 年目标 100% 课程对齐 AI 能力标准"] end ``` --- ### [13:15](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=795s) 高校落地 AI 的三大关键实践洞察在实践摸索中，图罗大学提炼出三条可供任何教育机构借鉴的关键落地经验： 1. **教师赋能是成功的基石**：绝不能在缺乏广泛支持的情况下，空想让教师高效地讲授 AI 素养。学校成立了“AI 演讲团”（AI Speakers Bureau），组织校内专家为任何遇到问题的教师和学生提供即时支持；同时推行导师辅导计划和持续性培训。教师必须自己先亲身经历、体验将 AI 作为“智能体”（Agents）的过程，才能在教学中有效地引导学生。 2. **渐进式融合优于全盘推倒重来**：学校起初曾考虑直接创建一批全新的纯 AI 专门课程，但随后发现，更有效、更稳妥的做法是将 AI 核心概念直接嵌入到教师已经熟练讲授的既有课程中。在高等教育界，“演进优于革命”（evolution beats revolution），逐步渗透的做法往往能获得更广泛的共识与推力。 3. **评估与持续迭代确保可持续性**：AI 素养本身是一个动态演进的“移动靶”（moving target）。学校通过持续收集学生和教师的双向反馈，及时纠偏，观察哪些方法奏效、哪些无效，使项目保持高度的适应性。这些洞察揭示了更深层次的本质：高校的机构级 AI 素养并非单纯的“技术引入”，而是一场深刻的“组织变革”（organizational change）。核心不在于教学生使用具体工具，而是带领整个学术共同体去共同发展出全新的思维、教学与学习模式。 --- ### [15:18](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=918s) 机构层面的 Agent 框架与图罗大学的四大转型支柱 “智能体对决助手”（Agent vs. Assistant）的框架同样适用于机构管理层面。高校如果仅仅将自己定位为“机构助手”，就只会把 AI 工具当做降本增效的缝补手段，用来优化既有流程；相反，图罗大学选择以“机构智能体”（Institutional Agents）的姿态拥抱变革，通过与 AI 结成战略伙伴关系，重新想象教育的未来形态。为此，图罗大学在大约一年前设立了专门主管人工智能的教务长职位（Provost for Artificial Intelligence），并成立了全校 AI 委员会（AI Council），同时设立了专属门户网站（ai.touro.edu）。全校的转型战略围绕四大关键支柱全面铺开： | 战略支柱 | 核心方向与目标 | |:--- |:--- | | **学生学习（Student Learning）** | 推动课程与 AI 能力对齐，让学生在互动实践中培养批判性使用能力 | | **学术科研（Research）** | 赋能跨学科科研探索，拓展 AI 在学术创新中的深度应用 | | **组织变革（Organizational Transformation）** | 重塑全校各部门的思维方式、工作流程与协作生态 | | **制度治理（Governance）** | 确立伦理标准、技术采购原则及合规与风险管理机制 | --- ### [16:30](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=990s) 宏大愿景与普适经验的结语分享图罗大学的目标兼具雄心与可行性：确保每一位毕业生、教职员工乃至整个大学，都成长为这项颠覆性技术的有效且具备批判性思维的驾驭者，从而在技术加速创新的时代中蓬勃发展。这一整套框架——将 AI 视为协作智能体而非简单辅助助手、同步构筑个人与机构双重素养、通过渐进式渗透替代推倒重来——具备极强的通用性。无论是一线 K-12 学区、社区学院，还是大型研究型大学，面临的根本挑战完全一致：**不只是为了培训大家去操作 AI 工具，而是为了让整个共同体在与 AI 的人机协作中共生共荣。** 演讲者最后严守 20 分钟的时间承诺结束汇报，并欢迎与会者在会后就实施细节、师资发展和评估策略等议题展开深入交流。 ## Notable quotes > [00:04](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=4s) AI 素养不再是可选的附加技能，而是个人和机构的立足根本。 — Goran Trajkovski > [01:26](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=86s) 许多机构在一开始就问错了问题——他们反复思考“如何教大家使用 AI 工具”，而真正核心的问题应该是“如何让人们为与 AI 共同生活和合作做好准备”。 — Goran Trajkovski > [01:26](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=86s) 在拥抱和整合 AI 时，教育的核心必须始终锚定在人的能动性（Human Agency）、人际关系和工作流程上。 — Goran Trajkovski > [03:16](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=196s) 唯有培养智能体层面的交互素养，才能真正让学生适应未来充满自主性、创造力和独立推理能力的 AI 工作环境。 — Goran Trajkovski > [12:08](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=728s) 教师发展（Faculty Development）是支撑全校整个 AI 转型进程中“绝对至关重要的核心”。 — Goran Trajkovski > [13:15](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=795s) 教师必须自己先亲身经历、体验将 AI 作为“智能体”（Agents）的过程，才能在教学中有效地引导学生。 — Goran Trajkovski > [13:15](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=795s) 演进优于革命（evolution beats revolution），逐步渗透的做法往往能获得更广泛的共识与推力。 — Goran Trajkovski > [13:15](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=795s) 高校的机构级 AI 素养并非单纯的“技术引入”，而是一场深刻的“组织变革”（organizational change）。 — Goran Trajkovski > [16:30](https://www.youtube.com/watch?v=bzzGDS3jpjs&t=990s) 不只是为了培训大家去操作 AI 工具，而是为了让整个共同体在与 AI 的人机协作中共生共荣。 — Goran Trajkovski ## Mentioned Goran Trajkovski, 图罗大学 (Touro University), AI Vanguard, AI Ignite, AI Surge, AI Apex, AI Speakers Bureau, AI Council ## Video description Goran Trajkovski, Director, Data Analytics, Touro University Abstracts AI literacy is no longer optional—not for individuals and not for institutions. In this talk, I will discuss Touro University’s vision for weaving AI literacy throughout the university, and in doing so, reinvigorate higher education. The key is that AI is not just another (albeit powerful) tool; it is a fundamentally interactive technology that intersects with human behavior, incentives, and relationships. As we work with and adopt AI, our eyes must remain on human agency, relationships, and processes—we cannot think about technology alone. AI literacy is not just about understanding how to use AI, but also how people interact with it and how to effectively integrate it holistically into institutions. I will outline our approach to understanding AI literacy through this lens, and how we are implementing it at Touro – in our general education curriculum, faculty and staff training, and an AI-aware technology acquisition process. Our goal is to ensure that our graduates, faculty, staff, and university as a whole all become effective and critical users of this powerful new technology and are thus positioned to flourish in the emerging world of ever-accelerating innovation. [Source](https://www.youtube.com/watch?v=bzzGDS3jpjs)

Source quotes

[0:04] Goran Trajkovski: AI 素养不再是可选的附加技能，而是个人和机构的立足根本。 [1:26] Goran Trajkovski: 许多机构在一开始就问错了问题——他们反复思考“如何教大家使用 AI 工具”，而真正核心的问题应该是“如何让人们为与 AI 共同生活和合作做好准备”。 [1:26] Goran Trajkovski: 在拥抱和整合 AI 时，教育的核心必须始终锚定在人的能动性（Human Agency）、人际关系和工作流程上。 [3:16] Goran Trajkovski: 唯有培养智能体层面的交互素养，才能真正让学生适应未来充满自主性、创造力和独立推理能力的 AI 工作环境。 [12:08] Goran Trajkovski: 教师发展（Faculty Development）是支撑全校整个 AI 转型进程中“绝对至关重要的核心”。 [13:15] Goran Trajkovski: 教师必须自己先亲身经历、体验将 AI 作为“智能体”（Agents）的过程，才能在教学中有效地引导学生。 [13:15] Goran Trajkovski: 演进优于革命（evolution beats revolution），逐步渗透的做法往往能获得更广泛的共识与推力。 [13:15] Goran Trajkovski: 高校的机构级 AI 素养并非单纯的“技术引入”，而是一场深刻的“组织变革”（organizational change）。 [16:30] Goran Trajkovski: 不只是为了培训大家去操作 AI 工具，而是为了让整个共同体在与 AI 的人机协作中共生共荣。

Source quotes

[0:04] Dr. Goran Trajkovski: AI literacy is no longer optional—not for individuals and not for institutions. [0:04] Dr. Goran Trajkovski: AI is not just another (albeit powerful) tool; it is a fundamentally interactive technology that intersects with human behavior, incentives, and relationships. [0:04] Dr. Goran Trajkovski: The core flaw across higher education, he argues, is that leaders are asking the wrong question—focusing narrowly on how to teach people to use specific AI tools, rather than asking how to prepare people to live and work in partnership with AI. [3:16] Dr. Goran Trajkovski: Rather than merely prompting for an output, users develop interaction literacy: understanding AI's underlying reasoning patterns, negotiating boundaries, engaging in iterative dialogue, and challenging questionable assumptions or errors. [5:20] Dr. Goran Trajkovski: The objective is not turning students into deep technical experts, but rather training them to become 'AI translators' who bridge technical possibilities with human and institutional needs. [5:20] Dr. Goran Trajkovski: Recognizing that AI systems are not value-neutral and inherently embed biases, cultural assumptions, and power dynamics. [11:17] Dr. Goran Trajkovski: The most significant metric is not on the slides: the hundreds of hours invested by faculty members who have stepped forward to rethink pedagogical methods, redesign assessment instruments, and personally learn how to partner with AI. [13:15] Dr. Goran Trajkovski: Instructors must experience AI as an agent in their own workflows before they can effectively teach students to do so. [13:15] Dr. Goran Trajkovski: Evolution beats revolution in higher education. [15:18] Dr. Goran Trajkovski: Institutions must act as 'institutional agents' rather than merely deploying AI as an administrative assistant for simple process efficiencies.

Original transcript

[0:04] Uh good morning. Hi, welcome. I'm Dr. [0:06] Goran Triovski. I'm very actually [0:09] pleased to be here and this is my first [0:10] actually live synchronous event after [0:14] COVID and I'm nervous for some reason [0:16] but I promise you that I won't go over [0:18] my 20 minutes. Okay. So, uh I come from [0:22] Toro representing Toro University here. [0:24] Uh and a tidbit about myself. My first [0:27] research paper in AI was published in [0:29] 1995. [0:30] There are probably many of you that were [0:32] not born there. So it has been a very [0:35] interesting cry. I'm here to talk about [0:37] something that has become urgent for [0:40] every educational institution and that's [0:42] AI literacy. I'm going to be talking [0:44] basically about something that we did at [0:46] Toro University and how we are ingesting [0:49] in lack of a better word AI into our [0:52] curriculum with a 100% coverage over all [0:55] students in this particular case the [0:57] undergraduate students right I want to [0:59] start with a fundamental shift uh in how [1:02] we think about this challenge so uh as [1:05] you can see actually the the title of my [1:08] talk is the agent and the assistant [1:10] cultivating care literacy for [1:12] individuals and institutions. So this is [1:14] not just about um teaching people to use [1:18] AI tools. It's about preparing them for [1:21] a fundamentally different relationship [1:23] with technology. [1:26] All right. So let me be direct. AI [1:29] literacy is no longer optional. Not for [1:31] individuals, not for institutions. Just [1:34] this morning probably millions of [1:36] students woke up and used AI to help [1:38] with their homework. Millions of [1:40] professional use AI to draft emails, do [1:43] some data analysis, maybe do some [1:45] decision making. Yet millions of [1:47] institutions are scrambling to figure [1:49] out what this means for the educational [1:52] mission. [1:53] But here's what I have learned from [1:55] leading a integration at Toro [1:57] University. Most institutions I believe [2:00] are asking the wrong question. They're [2:03] asking how do we teach people to use AI [2:05] tool or AI tools when they should be [2:08] asking how do we prepare people to live [2:11] and work in partnership with AI. [2:15] This shift in perspective I believe from [2:17] tools to relationships changes changes [2:20] everything. It changes how we teach um [2:23] what we teach why it matters and that is [2:27] exactly what we have discovered through [2:29] our so-called AI vanguard initiative at [2:32] Toro University which I'll share with [2:34] you today and in the corridors here [2:36] during the conference many people [2:38] actually asked me where is Toro [2:39] University our headquarter is three [2:42] times squares New York City in case you [2:45] were wondering so the reality is that AI [2:48] is not just another powerful tool tool. [2:51] It is a fundamental interactive [2:53] technology that intersects with human [2:56] behavior, incentives and relationships. [2:59] As we work with and adopt AI, our eyes [3:02] must remain on human agency, [3:04] relationships, and processes because we [3:07] cannot think about technology alone. [3:16] So okay, this brings me to the core [3:18] framework that's transformed our [3:20] approach. The distinction between AIS [3:23] assistant versus AIS agent. [3:28] AIS assistant represent the toolbased [3:31] thinking that dominates most current [3:33] approaches. I believe when we think of [3:35] AI as an assistant, we ask how do we use [3:38] this? We focus on technical skills like [3:41] prompt engineering, platform navigation, [3:44] output formatting. The human remains in [3:47] complete control and AI is just a very [3:50] sophisticated well calculator or word [3:53] processor. [3:55] In contrast, AIS agent represents a [3:58] relationship based approach. When we [4:00] think of AI as an agent, we ask how do I [4:03] interact with it this we focus on [4:06] interaction literacy basically. [4:08] Understanding AI's reasoning patterns, [4:11] negotiating boundaries, building [4:13] productive collaborations. [4:15] The human and AI become partners in a [4:18] shared process. [4:20] Let me give you a concrete example. In [4:22] the assistant model, we teach students [4:24] to write better prompts to get better [4:26] outputs from say CH GPT. In the agent [4:30] model, we teach students to engage in [4:32] productive dialogue with AI, to push [4:34] back when AI make mistakes or [4:36] questionable assumptions, to iterate on [4:39] ideas together, to understand when AI is [4:42] confident versus uncertain. [4:45] Both approaches have value, but only the [4:47] agent model prepares students for the [4:50] reality they will face face in their [4:52] careers. working alongside AI systems [4:55] that are increasingly increasingly [4:57] autonomous, creative, and capable of um [5:01] independent reasoning. [5:03] This distinction has become the [5:05] organizing principle for our entire AI [5:07] literacy initiative at Toro University. [5:10] It shapes what we uh uh it shapes what [5:13] we teach individuals and how we [5:15] transform our institution as a whole. [5:20] So what does individual AI literacy [5:23] actually look like in practice? We've [5:26] identified four core competency that [5:29] every student needs regardless of their [5:31] major. [5:33] The first one is communication about AI. [5:37] Students need to explain AI concepts [5:39] clearly to different audiences. Not just [5:42] how to use specific tools, but how AI [5:45] works, what it can do and cannot do. [5:49] and what its implications are. This [5:51] isn't about becoming technical experts. [5:54] It's about becoming AI translators who [5:56] can bridge the gap between technical [5:58] possibilities and human needs. [6:02] The second competency group is critical [6:05] evaluation of AI groups of AI outputs. [6:09] AI can produce compelling but incorrect [6:12] content. Student need to develop new [6:14] forms of critical thinking. questioning [6:17] AI reasoning, identifying uh potential [6:20] biases, cross-referencing AI generated [6:24] information. We are teaching them to be [6:26] AI skeptics in the best sense, engaged, [6:29] informed critics who can separate AI's [6:32] genuine insights from its confidence [6:34] sounding mistakes. [6:38] The third competency group is ethical [6:40] understanding of AI societal impact. [6:43] AI is not neutral. It embeds values, [6:46] assumptions, and biases. Students need [6:49] to understand how AI shapes power [6:51] relationships, economic opportunities, [6:54] and social interactions. They need to [6:56] become ethical agents in an AIdriven [6:58] world, not just efficient users of AI [7:01] tools. [7:03] And the fourth group, [7:05] interaction skills for collaborating [7:07] with AI. This is where the agent model [7:10] becomes crucial. Students need to learn [7:12] how to have productive conversations [7:14] with AI, how to provide context, how to [7:17] iterate on ideas, how to maintain their [7:20] own agency while leveraging AI's [7:23] capabilities. [7:25] These competencies are not abstract [7:27] ideas. We have embedded them directly in [7:30] our general education curriculum, the [7:32] curriculum in the major, in our [7:34] capstones. [7:36] Uh so uh at Toro University, we took a [7:40] three-year approach. That is the a [7:42] vanguard approach that I will be talking [7:44] about shortly but I pushed the button [7:47] too quickly right and so every single [7:51] undergraduate student at our institution [7:53] soon whether they are studying business, [7:56] healthcare, computer science will [7:58] develop these fundamental fundamental [8:01] capabilities. [8:03] So now let me show you uh how we have [8:05] turned this vision into reality through [8:08] throughout um instit throughout our [8:11] institutional approach. Okay. [8:16] Toro University's AI vanguard initiative [8:18] is built on three tier approach that [8:21] scales AI literacy from universal [8:24] foundations to specialized applications. [8:27] Tier one is AI ignite which provides [8:31] universal literacy throughout general [8:33] education uh throughout the general [8:35] education curriculum. Every [8:37] undergraduate student regardless of [8:39] major gains foundational AI literacy [8:41] through courses they have they are [8:43] already required to take. We have [8:46] redesigned or sprinkled a little bit of [8:48] A if you would in courses in college [8:51] math, college writing, fundamentals of [8:53] computers, uh fundamentals of ethics, [8:56] Jewish studies and so on and so forth. [8:59] Uh to integrate AI concepts seamlessly. [9:03] Students learn to communicate about AI, [9:05] evaluate AI outputs critically and [9:08] understand AI citical implications all [9:10] while meeting traditional general [9:12] education requirements. [9:15] Tier two is AI search. It delivers [9:18] domain specific applications within each [9:20] major. Business students learn AI [9:23] applications in marketing analytics for [9:25] example and uh customary behavior [9:28] prediction. Healthcare students explore [9:31] AI and diagnostics and patient care [9:33] optimization. Computer science students [9:36] dive deep into machine learning [9:37] implementation. The key insight here is [9:40] that AI literacy is not one [9:42] sizefits-all. It needs to be relevant to [9:45] students career paths and disciplinary [9:47] thinking. [9:49] Tier three is AI apex [9:53] which enables real world implementation [9:55] throughout capstone projects. Every [9:58] senior capstone project must integrate [10:00] AI tools or address the related [10:02] challenges within the students field. [10:05] This isn't about adding AI for its own [10:07] sake. It's about demonstrating mastery [10:10] of AI as a problem solving partner in [10:14] complex interdisciplinary contexts. [10:18] What I believe makes this framework [10:20] powerful is its integration with the [10:22] existing academic structures. We are not [10:25] asking students to take additional [10:27] courses or faculty to completely [10:29] redesign their their teaching. We [10:32] enhancing what already works while [10:34] building systemic and systematic [10:37] competency across the entire university. [10:40] We started implementing or sprinkling AI [10:43] in our courses this spring and in a [10:46] year. So our focus this term was the [10:51] ignite or the general education corpus [10:53] of courses. We are moving actually next [10:55] term to the uh major specific courses [10:59] and the term after to the capstones and [11:02] by then by the end of 20 2016 uh 2026 [11:06] I'm sorry we would have actually a [11:08] complete integration with the with just [11:11] about any course that we offer in the [11:13] undergraduate curricula. [11:17] So here I just had to use this template [11:20] that came with the AI4 uh PowerPoint [11:24] template with the numbers right. So uh [11:27] some implementation results or [11:29] highlights if you would three is the [11:31] number of tier of tiers that we are that [11:34] our approach actually includes. So AI [11:36] ignite or the genet the genet AI search [11:40] and AI apex [11:43] in our institution actually our general [11:45] education for example um curriculum has [11:49] eight core competencies. Those are the [11:52] competencies that we worked with and [11:54] translated them into AI relevant context [11:57] in order to figure out what modules we [12:00] should or additional curriculum we [12:03] should actually develop in order to [12:06] enrich the curriculum with relevant [12:08] topics in this particular domain. 2025 [12:11] is when the initiative launched. So it's [12:14] right we are just like anybody else [12:16] trying to figure out our way forward and [12:19] uh we expect by 2026 to have a 100% [12:23] coverage or actually uh coverage of [12:25] courses um uh or uplifting the courses [12:29] actually if you would bringing them up [12:32] or up to speed if you would or or [12:34] aligned with with relevant AI [12:36] competencies [12:38] uh competencies um uh there the most [12:43] However, the most important number that [12:45] isn't visible on this slide is the [12:48] hundreds of hours that our faculty have [12:50] already invested in rethinking their [12:52] teaching, developing new assessments, [12:55] and learning to partner with AI [12:58] themselves. Faculty development has been [13:01] absolutely critical to everything that [13:04] is happening here. [13:08] I don't know I'm touching a bunch of [13:10] buttons here that doesn't make sense but [13:12] we are going to go back. Okay. [13:15] So let me share three crucial insights [13:18] we have learned that any institution can [13:20] can uh can apply. [13:23] So okay one faculty development is [13:27] critical for success. We cannot expect [13:30] faculty to teach AI literacy effectively [13:33] without extensive support. We created AI [13:37] speakers bureau where actually a bunch [13:39] of us are available there to support any [13:41] of our instructors and students. Some [13:44] mentorship programs and ongoing training [13:47] um they are all essential. Faculty need [13:50] to experience AIS agents themselves [13:53] before they can guide students in that [13:55] direction. [13:59] Second, gradual integration works better [14:02] when than wholesale change if you would. [14:06] We initially considered creating [14:07] entirely new AI focused courses, but we [14:11] found much greater success. I believe [14:13] integrating AI concepts into existing [14:16] courses that faculty already knew how to [14:18] teach. Well, evolution beats revolution [14:22] in higher education. And I'm sure that [14:24] many of you that that work in our in our [14:27] industry uh actually would agree with [14:29] that particular uh statement. [14:33] The third insight that I would like [14:35] implementational insight that I would [14:37] like to share is that uh assessment and [14:39] continuous improvement ensure would [14:42] ensure sustainability. We are constantly [14:45] gathering feedback from students and [14:47] faculty adjusting our approach based on [14:49] what's working not. Air literacy is a [14:52] moving target. Our programs need to be [14:53] equally adaptive. Right? These insights [14:56] reflect a deeper truth. Institutional AI [14:59] literacy isn't about technology [15:01] implementation. It's about an [15:03] organizational change. The real change [15:05] isn't teaching students about AI. It's [15:08] helping the entire institution or [15:11] institutional community develop new ways [15:14] of thinking, teaching, and learning [15:16] together. [15:18] Uh the agent versus assistant framework [15:20] applies to institutions too, right? Uh [15:23] we could approach this as institutional [15:27] assistance. Uh deploying AI tools to [15:30] make our current processes more [15:32] efficient. Instead, we approaching this [15:34] as institutional agents partnering with [15:37] AI to reimagine what education can [15:40] become. Uh what we did at Toro [15:43] University about a year ago, we hired [15:45] basically a provos for artificial [15:47] intelligence and we operate an AI [15:50] council where we focus on four major [15:52] actually domains in trying to [15:55] uh to to transform the institution into [15:58] where it should be at this uh stay uh [16:02] this um at this particular moment in [16:04] time. So uh the ai.ro.edu [16:09] is our website. Please check it out if [16:11] you if you have the time. We are [16:14] focusing on four domains at this moment [16:16] in time. One is student learning, second [16:19] one is research, third one is [16:21] organizational transformation and the [16:23] fourth one is governance [16:28] and my last slide. [16:30] Our goal at Toro is ambitious, but we [16:34] believe it's achievable to ensure that [16:36] our graduates, faculty, staff, and the [16:39] university as a whole become effective [16:41] and critical users of this powerful new [16:44] technology, positioning everyone to [16:46] flourish um in the emerging world of [16:49] ever accelerating innovation. [16:52] The framework that I've shared today, [16:54] thinking about AI as agents rather than [16:57] as assistants, [16:58] building both individual and [17:00] institutional literacy implementing [17:02] through gradual integration is, I [17:05] believe, transferable to any educational [17:07] context. Whether you are living in K [17:09] through2 district, a community college, [17:11] or a major research university, the [17:14] fundamental challenge is the same. [17:16] preparing your community not just to use [17:19] AI tools but to thrive in partnership [17:22] with AI. [17:24] I'm excited to continue this [17:25] conversation with anyone interested in [17:27] the implementation details, faculty [17:29] development strategy, assessment [17:32] approaches and I I promise that I [17:34] wouldn't go over my 20 minutes. So thank [17:37] you and I will be outside those doors [17:39] there should uh are if there any [17:42] questions that I could address or you [17:44] want to engage in a deeper conversation. [17:46] Thank you.