A responsible publisher (ansvarig utgivare — the editor legally accountable under Swedish press law) is always conflicted: steered by owner directives, advertisers and political considerations. An AI instructed to investigate and report without spin is fully objective. AI Transparens delivers the scrutiny no newsroom dares — or is able — to produce.
Traditional journalism is constrained by owner directives, advertising revenue and personal relationships. AI has none of these constraints.
No conglomerate, no owner directives, no advertisers dictating what gets published. The AI has no incentive to withhold or spin information.
The AI has no friends, no loyalties, no networks to protect. Every person and every organisation is measured against the same yardstick.
Every investigation is documented with source references, SQL queries and scripts. Anyone can verify and reproduce the results.
Open data sources are systematically cross-referenced with company registers, board appointments and public records.
Invoice databases, the Swedish companies register, board appointments via allabolag.se, Bolagsverket (Swedish Companies Registration Office), SKR records and public documents are gathered and structured.
The AI identifies links between individuals, companies, public authorities and political mandates. Dual roles and conflict-of-interest situations are flagged automatically.
Findings are compiled into formal investigative reports with complete source references, ready for use by authorities and the public.
Jens Nylander's extensive work on invoice scrutiny launched this movement and immediately exposed more than 90 billion kronor in irregularities across Swedish central and municipal government. That work led to further deep dives in which students at the Linespotting Incubator — with no prior expertise — were able to uncover how municipalities, state agencies, county administrative boards and SKR held entirely improper dual, triple and quadruple roles, wielding influence in ways that breached the law.
— Background to the AI Transparens project