Format: 3 x 2-hour interactive workshops
Capacity: 25-30 journalists per session
Duration: 1 week
Time commitment: class time + 30-90 minutes of homework, total ~8 hours

SYLLABUS
Course Description
This track is designed for reporters who have a grasp of AI, spend a significant amount of their time covering technology, and want to go deeper. It will be an opportunity to clarify your understanding of technical concepts and think more expansively about how to cover the different facets of this fast-moving story. The course will require a dedicated time commitment: We will meet for a total of 6 hours in one week; there will be an additional hour of recommended homework between each session to get the most out of class time.
During the first session, we will cover the history of AI, the AI supply chain, and key technical concepts such as how to train a deep-learning model and the difference between supervised and unsupervised learning. We will also dive into basic data literacy skills, such as for investigating AI bias. On the second day, we will dig into what makes a good accountability story and how to report on governments and communities, including by documenting harms and embedding with affected populations. On the third day, we will dive deeper into more technical concepts related to generative AI (think: transformers, diffusion models, scaling laws), and how to report on companies, including by cultivating inside sources.
Finally, we will conclude with a pitch workshop for anyone interested in getting real-time feedback on an AI accountability reporting project.
How the Class Works
To make this course as inclusive as possible, we are conducting the trainings virtually. For you to get the most out of the course, and to support your peers in doing so as well, we ask that you treat it like an in-person workshop: close out your other work and commit to be 100% present and engaged during class time. If your internet connection and environment allows, we recommend you turn on your camera.
Learning Outcomes
Through this course, participants will learn:
- background knowledge on the history of AI to understand its latest developments
- a clearer understanding of how AI works and how to better cover the multiple parts that make up its supply chain
- how to resist AI hype, and how to identify and cover the most important dangers, failures, and real world impacts of AI
- practical reporting methods to report on AI, including basic spreadsheets, public records requests, and strategies on approaching sources working at tech companies
- techniques for investigating bias in automated systems and tracking misinformation
- how to bulletproof your reporting with the help of experts and fact checkers
- how to formulate clear pitches around AI, including for accountability stories
Structure of the Course
Subject to change based on the needs of workshop attendees.
Day 1: Foundations
Preamble (15 minutes)
- Introduction to the instruction team, structure & goals of the course
- Introductions to your classmates
What is AI (45 minutes)
- History of AI and the AI lifecycle (inputs / outputs)
- Intro to deep learning: supervised, unsupervised, and reinforcement learning; different neural network architectures
- Build a machine learning model (no coding necessary!) - interactive
Break (5 minutes)
Data skills (40 minutes)
- Basic data literacy for investigating tech
- Investigating bias - interactive
Close (5 minutes)
- Summary, questions/requests
- Homework for next week
Homework (60 minutes)
- Watch neural networks videos
But what is a neural network? | Chapter 1, Deep learning
Gradient descent, how neural networks learn | Chapter 2, Deep learning - Think about a pitch for an AI accountability story
Day 2: Reporting on governments and communities
Themes for the day (2 minutes)
AI reporting through an accountability lens (55 minutes)
- Anatomy of an accountability story - interactive
- Social media and misinformation
- Documenting harm
Break (5 minutes)
Local on-the-ground reporting (50 minutes)
- Navigating government agencies + FOIA
- How I reported this: Embedding within communities + Q&A - interactive
Close (5 minutes)
- Summary, questions/requests
- Homework for next week
Homework (60 minutes)
- Read article about transformers
Generative AI exists because of the transformer - Refine your pitch for an AI accountability story
Day 3: Reporting on companies
Themes for the day (2 minutes)
Interrogating the generative AI wave (50 minutes)
- Introduction to Generative AI: transformers, diffusion models, scaling laws, reinforcement learning from human feedback, hallucinations
- How I reported this: Cultivating sources within companies + Q&A - interactive
Break (5 minutes)
Dealing with companies (30 minutes)
- A guide to AI corporate jargon - interactive
- How to bulletproof your reporting and analysis
Pitch workshop (20 minutes)
Close (10 minutes)
- Summary and feedback

ALL TRACK 2 TRAININGS

AI SPOTLIGHT SERIES
AI Spotlight Training for Southeastern European Journalists
This training, hosted in partnership with iMEdD, will be in person, held in English, and targeted towards journalists in Albania, Bosnia and Herzegovina, Bulgaria, Croatia, Greece, Kosovo, Montenegro, North Macedonia, Romania, Serbia, Slovenia, and Turkey.