In this skills training, we will focus on how to use AI chatbots effectively for tasks throughout the research process. We will cover how to customize AI Chatbot settings and practice the art of writing effective prompts. We will also briefly examine how Large Language Models (the models behind modern chatbots) work, get an overview of different models and their strengths, and discuss the most important ethical questions around using AI chatbots in research.
- Workshop developed by: Dr. Hannah Metzler & Konstantin Hebenstreit, MSc.
- Introduction & Ethics shorter version
- Introduction to LLMs
- Ethics
- Chatbot essentials
- Advanced prompting strategies & Hallucinations
- Customizing Chatbots: general and system prompts, memory, CustomGPTs/Gems
- LLM Research Reports (Deep Research)
- Scientific tools based on LLMs: NotebookLM, Perplexity, Elicit
- Scientific writing with chatbots
- Collection of researcher use cases
- Exercises: Coding & coaching
- Resources
- April 29 2025: Skills training at Ludwig Boltzmann Gesellschaft Career Center, Vienna: Welcome & course overview
- January 9, 2025: Austrian Academy of Sciences, Austrian Archaeological Institute : Welcome & course overview
- December 4, 2024: Karl Landsteiner Privatuniversität Krems, Zentrum für Transitionspsychiatrie: Welcome & course overview
- October 29 2024: Skills training at Ludwig Boltzmann Gesellschaft Career Center, Vienna: Welcome & course overview
For the workshop exercises in the LLM workshop, we recommend having access to a pro/plus version of one of the best chat bots, starting from the day of the workshop. It's around €25/month, with the option to cancel before the renewal date every month. In all cases, you can cancel your subscription within the first month, and will not have to pay for a second month. You could also choose to switch from one model to another after the first month. Please make sure you have access to one of the below options before the workshop.
-
Option 1: ChatGPT Plus: most multimodal features
- Sign up at chat.openai.com.
- Upgrade to Plus (under your account symbol in the top right corner).
- To turn off model training on your data: Go to Settings → Data Controls → Improve model for everyone: off.
- Download the ChatGPT app on your phone. If you're using a MacBook, install it on your laptop as well.
-
Option 2: Claude If you primarily use LLMs for writing and value a strong writing style, Claude.ai is a good alternative. Claude.ai prioritizes privacy and doesn't use your input for model training (except for safety concerns).
- Sign up under claude.ai
- Click Upgrade at the top center
- Subscribe to Pro
- (Claudet is a voice extension you can only use if you buy API access to Claude. Not for this workshop.)
-
Option 3: Google Gemini Pro (Free Trial) Google offers a free trial of Gemini Pro for one month. If you cancel before the trial ends, there's no cost.
- Sign up under https://gemini.google.com/
- You can choose to turn off the "Gemini Apps Activity" setting. This also prevents any data from being saved. There is no way to only turn off model training on your data while keeping earlier chats in Gemini. You can choose to save your Activity for 3, 18 or 36 months, before it gets deleted.
Looking forward to seeing you at the workshop!