For everyone

A century ago, when canned foods and bottled medicines first filled grocery shelves, ordinary shoppers weren't chemists. They couldn't test a tin of milk in a home laboratory. What changed the food supply wasn't every parent learning organic chemistry, it was families, co-ops, and town clubs insisting on honest ingredient labels, clean kitchens, and inspectors who answered to the public.

You don't need to know how to train a neural network to insist on fair ground rules for AI. If you use a phone, pay a medical bill, belong to a credit union or union local, or raise children, you already have real leverage.

Five things one person or household can do this week

You do not need technical expertise to insist on fair ground rules. Everyday habits (like loaning narrow permissions, keeping intimate notes on local devices, and asking local officials for human appeal routes) keep people firmly in charge.

1. Treat permissions like house keys, not confetti. When a new AI app asks for access to your email, calendar, contacts, and payment cards at once, pause first. Ask yourself: would I hand a stranger my front-door key and my wallet just to book a haircut? Prefer tools that let you grant one narrow permission at a time and turn it off when the errand is done. Read our guide on AI agents should hold permissions that expire.

2. Ask your tools for a receipt you can keep. If an assistant summarises a contract, checks a bill, or sends a message on your behalf, check whether it gives you a plain, dated record of what it read and what it did. If a tool hides its steps or refuses to let you export your own notes, treat it with the same caution you would show a shop that refuses to print a receipt. Read our guide on how to tell if an AI assistant works for you.

3. Keep intimate notes on devices you control. You don't have to upload private family health questions, journals, or household budgets to an ad-funded cloud service. Increasingly, small open models can run right on an ordinary laptop or phone with the internet unplugged, like writing in a notebook on your kitchen table instead of on a public bulletin board.

4. Check the fine print before you sign or pay. Medical bills, rental leases, and service contracts are often generated by software that counts on you being too busy to check the rules. Using a local, plain-language assistant to cross-check an itemized hospital bill against public billing rules, or to highlight auto-renewal traps in a contract, turns the tables back in your favour. Read our guides on reading a medical bill and checking a contract.

5. Ask your elected representative one specific question. Politicians receive thousands of form letters, which are easy to ignore. A short, specific note from a constituent gets read. Ask your city councillor, school trustee, or MP two simple questions. Does our municipality keep a public list of automated systems used in public decisions? And can residents appeal to a human with the power to reverse an error?

For co-ops, clinics, unions, newsrooms, and community groups

Individual choices matter, but organised neighbours have ten times the bargaining power of a lone consumer. When thirty independent clinics, a credit union, a teachers' local, or a housing co-op act together, vendors listen.

Pool your buying power without giving up your privacy. Small businesses, independent pharmacies, and rural clinics often overpay for supplies and software because they bargain alone. Using privacy-preserving order pooling, local organisations can combine their purchasing volume to get fair wholesale terms without ever exposing their private books or customer lists to competitors. Read our explainer on pooling orders without showing your books.

Protect craft while dropping the paperwork chores. In workplaces, the worst use of AI is turning skilled nurses, teachers, mechanics, or journalists into hurried clerks who rubber-stamp machine outputs. Unions and professional associations can bargain for a clear boundary: use software to clear away repetitive form-filling and scheduling chores, while keeping human judgment, mentorship, and final sign-off in human hands. Read our essay on keeping the craft and dropping the chores.

Pair digital polls with real kitchen-table conversation. When a co-op, union, or neighbourhood association needs to set priorities, avoid outrage-driven comment threads or simple thumbs-up polls where the loudest faction wins every time. Combine small, representative discussion tables with bounded voice budgets so members can show how deeply they care about a shared need. Keep human facilitators in charge of community deliberation, and welcome everyday speech instead of bureaucratic jargon.^1,2,3,6,7,8 Read our essay on hearing care, not just volume.

Challenge black-box screening in housing and work. When landlords, schools, or employers use automated screening scores or algorithmic rent-setting tools, tenant associations and worker groups can reject generic compliance badges. Demand to see the exact data used, the receipts proving accuracy, and a human reviewer with authority to fix errors. Civil rights and consumer protection rules do not disappear just because a decision was made by a spreadsheet or a model.^4,9

Document real-world failures in the open. When an automated system wrongly cuts off a neighbour's benefit, flags a student's honest essay, or denies a valid claim, write down the details. Record the date, the notice given, and what happened when you tried to appeal. Submitting real cases to public registries like the AI Incident Database helps researchers and regulators stop the same failure from repeating in the next town.⁵

You can also explore Shared Horizon to see how over three hundred organisations score on human oversight today. From tech giants to co-ops, see what practical moves will help each group move forward.

Write to us at contact@collectiveintelligencenetwork.org. A person reads every message, because keeping people in charge of AI starts with listening to the people it affects.

For researchers

Democratic deliberation and civic coordination build on computational social choice. This includes the Method of Equal Shares for proportional budgeting (Peters & Skowron, 2020) and preference clustering (Yang & Bachmann, 2025²; Fishkin, 2009²). Automated rule discovery (Thach et al., 2026³), algorithmic shortlisting (Zambrano et al., 2025³), and fair exposure ranking (Hausladen, 2026³) balance social welfare with fairness for underrepresented cohorts.

In citizen assemblies, idea lineage tracking (Poole-Dayan et al., 2025⁶), proposal stability audits across input sequence permutations (Mahajan et al., 2026⁶), and deliberative scaffolding (De Liddo et al., 2026⁶) prevent summarisation erasure and expose silent choices in automated synthesis. This addresses findings that synthetic model groups produce one-third the perspective diversity of human assemblies (Flechtner, 2026⁶). Representative surveys reveal an AI penalty in civic engagement (Jungherr & Rauchfleisch, 2025⁷). Automated facilitation sharply decreases participation willingness and perceived quality, proving software must remain subordinate to human moderators. Linguistic inclusion research (Sharoff, 2026⁸) further shows that default model registers risk privileging bureaucratic communication styles, requiring deliberate socio-demographic scaffolding to empower marginalized participants.

Stakeholder-specific transparency research (Smith et al., 2026⁹) demonstrates that conventional static disclosures fail end-users, requiring pluralistic, situationally-constructed transparency infrastructures. In housing and algorithmic accountability, we draw on empirical civil enforcement research (Chien et al., 2025⁴). Model multiplicity doctrines (Black et al., 2024⁴) require entities deploying predictive filters to search affirmatively for less discriminatory algorithms.

1. OECD, Innovative Citizen Participation and New Democratic Institutions: Catching the Deliberative Wave (Paris: OECD Publishing, 2020); Michiel A. Bakker et al., "AI can help humans find common ground in democratic deliberation," Science, Vol. 386, No. 6720 (2024), pp. 310–315.

2. Joshua C. Yang and Fynn Bachmann, "Bridging Voting and Deliberation with Algorithms: Field Insights from vTaiwan and Kultur Komitee," ACM FAccT (2025); James S. Fishkin, When the People Speak: Deliberative Democracy and Public Consultation (Oxford University Press, 2009).

3. Nguyen Thach, Xingchen Sha, and Hau Chan, "Large Language Models for Designing Participatory Budgeting Rules," AAMAS (2026); Juan Ignacio Zambrano et al., "Algorithmic Shortlisting in Participatory Budgeting," arXiv:2508.06577 (2025); and Carina I. Hausladen, "Fair Feed Ranking for Participatory Budgeting," ACM GoodIT (2026).

4. Colleen V. Chien, Anna Zink, and Irene Y. Chen, "Bridging Research Gaps Between Academic Research and Legal Investigations of Algorithmic Discrimination," arXiv:2508.14954 (2025); Emily Black, John Logan Koepke, Pauline T. Kim, Solon Barocas, and Mingwei Hsu, "The Legal Duty to Search for Less Discriminatory Algorithms," arXiv:2406.06817 (2024); and U.S. Department of Justice, "Justice Department Sues RealPage for Algorithmic Pricing Scheme," (August 23, 2024).

5. AI Incident Database, "Submit New Incident Report"; Sean McGregor, "Preventing Repeated Real World AI Failures by Cataloging Incidents: The AI Incident Database," AAAI, Vol. 35, No. 17 (2021), pp. 15458–15463.

6. Elinor Poole-Dayan, Deb Roy, and Jad Kabbara, "An AI-Powered Framework for Analyzing Collective Idea Evolution in Deliberative Assemblies," MIT Center for Constructive Communication, arXiv:2509.12577 (2025); Sachit Mahajan, Javier Argota Sánchez-Vaquerizo, Arthur Capozzi, and Dirk Helbing, "From Opinions to Commitments: Unstable Assignments and Silent Choices in AI-Mediated Proposals," ETH Zurich, arXiv:2611.05135 (2026); Anna De Liddo, Lucas Anastasiou, and Simon Buckingham Shum, "Human/AI Collective Intelligence for Deliberative Democracy," Handbook of Democracy in the Age of AI, arXiv:2603.16260 (2026); and Maurice Flechtner, "The Deliberative Deficit: An Empirical Critique of LLMs in Democratic Discourse," arXiv:2608.10186 (2026).

7. Andreas Jungherr and Adrian Rauchfleisch, "Artificial Intelligence in Deliberation: The AI Penalty and the Emergence of a New Deliberative Divide," Government Information Quarterly (2025), arXiv:2503.07690.

8. Serge Sharoff, "The Almost Intelligent Revolution: Options for Scaling Up Deliberation and Empowering People with AI," arXiv:2606.19864 (2026).

9. Genevieve Smith, Sahiba Chopra, Natalia Luka, and Min Kyung Lee, "Whose Transparency is it Anyway? Understanding Generative AI Transparency Needs and Gaps Across Stakeholders," arXiv:2611.05549 (2026); and José Ramón Enríquez, Jiaxin Pei, and Alex Pentland, "AI Should Facilitate Democratic Deliberation at Scale," ICML, arXiv:2609.20059 (2026).