About FairByDesign
I follow the systems behind the claims.
FairByDesign is an independent reporting and research operation examining how AI is changing power, work, security, institutions and everyday life. We follow the evidence, investigate the systems behind the claims, and build practical tools and research for people who need to understand what is actually happening.
A reporter’s note
Why I do this work
I came to AI through systems: data, security, governance, architecture, and the uncomfortable space between what an organization says a system does and what the system actually does. My background combines academic research in cybersecurity law with real-world governance work, and FairByDesign grew out of that gap.
Security is where I started, and it is still where I look first. AI systems are no longer just answering questions. They are being handed credentials, tools and the authority to act. Every new capability is also a new attack surface, and when something goes wrong it is rarely just a breach. It is a decision taken on someone’s behalf that nobody can fully reconstruct or reverse.
That is also why sovereign AI matters to me. More of our institutions, companies and public services now depend on models, data and infrastructure they do not own, cannot inspect and could lose access to overnight. Access you can buy is not the same as access you control. Who holds that control, and what happens when it shifts, is one of the defining accountability questions of this decade.
And underneath all of it is a simple conviction: technology should be fair by design. AI already shapes work, credit, healthcare, education and family life, usually long before the people affected are given clear language for how those systems work. Socially conscious technology is not a slogan. It means building, buying and governing systems as if the people on the receiving end matter, because they do.
So I report on the machinery. Sometimes that means reading legislation and following who asked for a change. Sometimes it means examining a company’s public claim against the architecture required to make that claim true. Sometimes it means building a test and seeing what happens.
The point isn’t to predict whether AI will save us or destroy us. The point is to find out what is happening while there is still time to do something about it.
“Access you can buy is not the same as access you control.”The question behind FairByDesign’s sovereign AI reporting
What we watch
One beat, followed from policy to architecture to security to evidence to human consequence.
Control of AI agents
What AI systems are allowed to do, who gave them that authority, and whether anyone can actually stop, audit or reverse what they did.
Who controls the stack
Dependence on models, data, cloud and vendors that organizations and public institutions cannot inspect, keep or replace.
Evidence, not assertion
Whether governance claims, standards and product promises survive contact with running systems, and who is responsible when they don’t.
The people on the receiving end
How AI decisions land on work, rights, access and opportunity, and what fair-by-design technology would require instead.
We show our work
FairByDesign doesn’t just tell you what happened. It lets you inspect how we know.
- Sources and documentsPrimary records linked wherever we can, with full references on substantive pieces.
- What we know, and what we don’tVerified findings are kept apart from analysis, disputed claims and open questions.
- MethodHow a finding was reached, including Evidence Lab protocols where we test a claim ourselves.
- Right of replyWe seek responses from the people and organizations we report on where appropriate.
- Corrections and updatesMaterial changes are dated and noted on the story.
- AI-use disclosureEverything is researched, written and reviewed by FairByDesign. Claude, an AI model made by Anthropic, is used for editing. It never manufactures evidence.
Independence & funding
FairByDesign is independent of the vendors and institutions it examines.
Who pays for the work
Reporting is funded by readers who subscribe to Mercury Intelligence, community supporters, public-interest underwriting from institutions, our books and training, and clearly labeled advertising and sponsorship.
Underwriters and supporters fund the work. They do not choose findings, review stories before publication or receive favorable coverage.
Advertising and sponsorship
We welcome advertising and sponsors, from tools and products we think readers may find useful to podcast, video and YouTube sponsorships, as long as there is absolutely no conflict of interest with our independent reporting.
Sponsored material is always labeled as such and kept separate from investigations. Sponsors have no say over what we cover or conclude.
How FairByDesign fits together
The reporting is the front door. Everything else is a way to go further.
Investigations, the free Mercury Brief, Evidence Lab and the Ethics Engine podcast.
Latest reportingThe Mercury BriefEvidence LabEthics EngineMercury Intelligence: full investigations, evidence notes, frameworks and the research archive.
Mercury IntelligenceField guides and playbooks born from the reporting.
Books & field guidesGet in touch
Tips, documents & corrections
Have firsthand evidence, or think we got something wrong? Email info@fairbydesign.org.
For anything sensitive, write to viktoriabakos@proton.me. Proton Mail is end-to-end encrypted when you write from a Proton account of your own, so that gives the most protection. Please don’t send documents that could identify you until we’ve agreed how to handle them.
Press & speaking
Interviews, panels and guest lectures: see the press page or email info@fairbydesign.org with your outlet, deadline and topic.