FairByDesign
Independent reporting on AI, power, security and accountability.
We investigate what AI companies, governments and institutions say their systems can do, what the evidence shows, and what happens when those systems meet the real world.
Latest investigation
The story we are following
-
Who Benefits From Slowing Down AI? Sanders, the Frontier Labs and the UN in One Week
A bill to ban superintelligence, two frontier model launches on the same day, and a UN that wants a brake it cannot build. An open-source reading of who pays first, and who benefits.
This week
Reporting, testing and research
-
Mercury Brief: The Investigation Is Funded by People Who Want It to Exist
This week FairByDesign launches the 2027 Independence Fund, a ten-day Evidence Chain series and the Who Benefits dialogue. Here is what is coming, how to give,…
The Agent Control Gate
Can runtime controls stop unauthorized AI-agent tool actions? Four control conditions, eight scenarios, a 320-run target.
In progress · September 2026
-
Is an AI Policy Enough?
An airline’s chatbot invented a refund rule, and the airline argued it was not responsible. A policy tells you what should happen. It cannot tell you…
Evidence Lab
We don’t just report what systems claim. We test what they do.
Evidence Lab builds controlled test environments, instruments them, attacks them and publishes enough of the method for others to reproduce the result. Policies, standards and product claims do not prove that a control works.
- ReportWhat the claim says
- TestA controlled, reproducible experiment
- FindWhat actually happened
- ExplainWhat it means and what it doesn’t
- DocumentMethod, data and limits, in public
Follow the reporting
The Mercury Brief
The free weekly briefing from FairByDesign: what happened, what we found, what matters and what we’re investigating next.
No hype roundups. No content treadmill. Just the reporting worth following.
Go deeper
Mercury Intelligence
The public publication tells you what we found. Mercury Intelligence lets you go behind the finding: full investigations, research papers, evidence notes, technical analysis, frameworks, the research archive and the AI Retooling Library.
If AI is changing your profession, it also helps you see which skills transfer, where the gaps are and what to learn next.
The library
Field guides born from the reporting
Public AI RealityThe broad entry point: what AI is already changing
Operational AI ReadinessFree to read: oversight for AI already running in live workflows
EU AI Act for Builders and OperatorsObligations translated into concrete decisions
Enterprise AI Agent Deployment PlaybookSecond Edition: evidence, release gates and go/no-go for AI agentsThe Ethics Engine
Podcast & video
FairByDesign’s public-interest reporting and conversation series: investigations, interviews and public explainers.
For organizations
When you need to act on it
Sovereign AI Security & ControlWhat an organization can do about control, access and dependence.
Enterprise AI Agent DeploymentEvidence, release gates and go/no-go for AI agents.Watch the 6-minute demo · Module details
Executive AI Decision ProgramA private, self-paced decision program for leaders.
Training, workshops & clinicsTeam support from understanding to operating practice.
Skills & ToolsThe AI Retooling Library and open-source tools.
App HubThe FairByDesign App Hub.
Support the reporting
Keep independent reporting possible
FairByDesign is independent of the vendors and institutions it examines. Reader and institutional support funds the reporting, research, public explanations and Evidence Lab work that remain available outside the commercial layer.
- FollowThe free Mercury Brief
- FundCommunity support
- UnderwritePublic-interest underwriting
- Go deeperMercury Intelligence
About the reporter
I follow the systems behind the claims.
Viktoria Bakos reports on AI from inside the machinery: data, security, governance, architecture and the space between what an organization says a system does and what it actually does.