Field Guide
Symposium Research & Design Tips
Projects to learn from, and questions to keep asking, as you research your system and design your exhibit.
Projects to explore
Public education and advocacy projects. Look at how each one gets a visitor to do something, not just read, and what you would borrow or change.
A game where you play a hiring manager and see how biased training data shapes automated hiring.
An interactive walk-through of how face-analysis AI judges the person in front of the camera.
A pop-up exhibition from Tactical Tech that makes data and technology tangible for visitors.
A “nutrition label” for datasets that shows what is inside and what its limits are.
A searchable collection of reports of harms and near-harms caused by AI systems.
A set of cards for spotting and addressing blindspots in AI systems (MIT Media Lab).
Research on the social implications of AI and the concentration of power in the tech industry.
Research, art, and advocacy that highlight the harms and biases of AI systems.
A curriculum project for teaching young learners about AI and ethics (MIT Media Lab).
A hands-on activity where learners train a model and see how its training data shapes its behavior.
An interactive game about AI ethics, hosted on itch.io.
Free, more accurate and diverse images for representing AI, with guidance on choosing them.
Where to find controversial systems
A system is worth investigating when people disagree about it in public. Look for these signals:
- Lawsuits, settlements, or regulator investigations
- Journalists or researchers who audited it and found uneven results
- Bans, moratoria, or new laws aimed at it
- Workers, artists, or communities organizing against it (strikes, open letters, protests)
- A company that paused, changed, or retired the product after criticism
- The maker and its critics describing the same system in very different ways
Search tip: pair a technology with words like lawsuit, audit, investigation, ban, moratorium, or settlement. The AI Incident Database above is a good first stop, and our own Examples show what a well-documented case looks like.
Places to look
AI incidents and hazards drawn from news coverage worldwide, searchable by sector and type of harm.
A database of court cases involving AI, useful for finding systems people are suing over.
A tracker of copyright lawsuits against AI companies.
A searchable map of surveillance technology used by U.S. law enforcement agencies.
A curated list of troubling or harmful uses of AI.
Research and reporting on automated decision-making and its effects on people.
Investigative journalism about how technology is used on people.
Investigations of algorithms used in criminal justice and other high-stakes decisions.
Independent tech journalism covering surveillance, platforms, and AI.
MIT Technology Review: Artificial Intelligence
Reporting and analysis on new AI systems and their consequences.
Analysis of technology policy, regulation, and governance.
Cases and reports on surveillance and algorithmic decision-making.
An annual report with data on AI development, policy, and incidents.
What makes a good exhibit piece?
Whatever system you choose, help visitors see what is happening beneath the surface:
- Representation: What does the system receive, and how is the world turned into data?
- Procedure: What does the system actually do with that information?
- Priorities: What is it trying to get right, and who decided what counts as success?
- Impacts: What changes when the system is used, and for whom?
- History: How is this similar to technologies, systems, or controversies that came before? What is genuinely new?
Across all five, keep asking: How are we deciding what should happen?
Keep the benefits in view
You can be critical of many AI systems and still engage with what they offer. AI isn't just one thing.
- Accessibility: captioning, speech-to-text, image descriptions, translation, assistive communication
- Healthcare: image analysis, clinical decision support, drug discovery, administrative burden reduction
- Science: protein structure, climate modeling, astronomy, materials discovery
- Education: tutoring, feedback, language support, practice generation
- Creative work: brainstorming, prototyping, editing, coding assistance
- Safety and maintenance: anomaly detection, fraud detection, predictive maintenance
- Everyday convenience: navigation, spam filtering, search, recommendations