In this introduction, we will try to pin down the nebulous term “AI” by looking at historical precedents and the component parts, including Large-Language Models. We will survey the space of current commercial AI, with what are often referred to as “frontier” or “foundation” models, and begin our work with Claude by exploring the interface and settings and completing some introductory exercises.

By the end of the session, every participant will have a working Claude account configured to their preferences and have built a small ELIZA-style chatbot together as a Claude Artifact.

NEH Workshop 1 — Wednesday, May 13, 10 AM – noon, CHDR

Streamed and recorded. Open to UCF faculty, graduate students, and the larger arts and humanities community.

Open the slide deck →

What to Bring

Pre-Workshop Reading

Light prep before Wednesday so we can go deeper together:

Optional, if you want to be a step ahead:

Session Outline (120 minutes)

The workshop is built in four acts. Slide deck linked above; this outline maps the flow.

  1. ELIZA and the ELIZA effect. Weizenbaum’s 1966 chatbot, the secretary who asked him to leave the room, the two roads from MIT (Weizenbaum’s lifelong critique vs. Minsky’s AI Lab and Pentagon funding). We try the masswerk.at web ELIZA live and notice what our brains do.
  2. AI, LLMs, and how they actually work. AI as a field, ML as a subfield, deep learning as a method, LLMs as a particular use of deep-learning transformers. Tokens (try the OpenAI Tokenizer), pretraining vs. inference, RLHF and reasoning layers, the Opus / Sonnet / Haiku family.
  3. The ChatGPT moment. What already existed before November 30, 2022 — the GPT-3 API (June 2020), GitHub Copilot (June 2021). What the chatbot interface added: access. Why “the chatbot was the wrapping” is a useful sentence to have in your back pocket.
  4. The current landscape and the agentic horizon. A short survey of what people mean by “frontier” and “foundation” models — Claude (Anthropic), GPT-5 (OpenAI), Gemini (Google), Llama (Meta open weights). Where the conversation is shifting now: from did the chatbot write this? to did the agent do the work? Reading anchor: the Chronicle on agentic AI in higher ed.

The hands-on portion runs through the second half:

  1. Chatbot tour: five conversations. Same prompt, five tools, in this order:
    1. Google Gemini — free, fast, integrated with Google’s ecosystem; the model many of our students already use.
    2. DeepSeek running locally through Ollama — an open-weights model on a personal machine, no API call, no data leaving the room. The “what does local feel like” demo.
    3. ChatGPT — the chatbot that defined the genre, OpenAI’s flagship.
    4. UCF Copilot — the institutional option most of you already have. Microsoft’s wrapper around an OpenAI model, with UCF’s data-privacy framing on top. This is the only tool UCF currently supports or allows for use with restricted data. See UCF AI Resources for the current policy.
    5. Claude — Anthropic’s flagship; the tool we use for the rest of the series. We will be moving back and forth between these throughout the workshop while prioritizing Claude, to build an understanding of the different tools available.
  2. Why Claude / why Anthropic.
    • Tool surface. Projects + Artifacts + Code Web + CLI is currently the broadest set of non-coder-accessible tools across any provider. The conceptual framework transfers, but right now the build path is best on Claude.
    • Safety + interpretability research culture. Anthropic publishes more on alignment and model behavior than its peers; for a humanities audience, that publication record is part of what we’re choosing.
    • The series is designed so the framework transfers if you choose a different tool — and I’ll point at where it would.
  3. Setting up Claude. Subscription confirmation. Settings tour (model selection, conversation history, custom instructions, voice mode if available, Artifacts toggle). Style preferences. Why these choices matter for the rest of the series.
  4. Build our own ELIZA in a Claude Artifact (group exercise). Prompt Claude together to construct a Rogerian-style chatbot Artifact in the spirit of Weizenbaum’s 1966 original — but with stylistic latitude. Try it. Compare to the masswerk ELIZA. Observe where the ELIZA effect kicks in, sixty years on.

Core Exercise

Build our own ELIZA — together, in Claude Artifacts. This is the workshop’s closing exercise and the asynchronous equivalent for those who cannot attend.

Pick a style, tone, and subject of your own based on your own interests — something connected to your discipline, your teaching, or a corner of your life you’d be willing to talk to a chatbot about. The pattern is the constraint; the personality and the topic are yours.

  1. Open Claude and start a new chat.
  2. Prompt: “Build a Claude Artifact: a chatbot in the spirit of Weizenbaum’s 1966 ELIZA — pattern-match the user’s input and reflect it back as questions. Don’t pretend to know more than the pattern.” Then add your own style, tone, and subject.
  3. Open the Artifact in the side panel. Talk to it the way you talked to the masswerk ELIZA.
  4. Compare. What does Claude do that pattern-matching can’t? What does it do that pattern-matching also did? Where does the ELIZA effect kick in this time?
  5. Save the Artifact URL — bring it to the W2 async discussion.

A Frontier-Model Quick Reference

A pocket survey of the major commercial “frontier” or “foundation” models as of Spring 2026. The names move fast; the players are roughly stable.

Maker Family Used in this series
Anthropic Claude (Opus, Sonnet, Haiku) All workshops
OpenAI GPT (5, 4o), Images W1 chatbot tour (ChatGPT); W6 multimodal (OpenAI Images)
Google Gemini (Pro, Flash, Nano Banana) W1 chatbot tour (Gemini); W6 multimodal (Nano Banana)
DeepSeek (open weights) DeepSeek V3 / R1 W1 chatbot tour (local via Ollama); W11 local-models discussion
Microsoft Copilot (wraps OpenAI) W1 chatbot tour (UCF Copilot)
Meta Llama (open weights) W11 local-models discussion
xAI / Mistral / others Grok, Mistral Referenced

We use Claude across the series for one practical reason: it is the model whose Projects, Artifacts, and Code Web tools currently make it easiest for non-coders to build something they can share. The conceptual framework transfers across providers — and W1’s chatbot tour is the moment we make that transfer-ability literal.

Pedagogical Note

We do not begin with awe and we do not begin with despair. We begin with sobriety. The students sitting in our courses this fall did not ask for this technology and many of them are anxious about it. One useful thing we can do as instructors is to know where these systems came from, what an LLM actually is, what reasoning and human-feedback layers add on top, and refuse to gesture vaguely. After this session you should be able to explain to a colleague: here is the ELIZA effect, here is what an LLM does, here is what changed on November 30, 2022, and here is what “agentic” means now.

Cross-references