Will AI Become Your Delegate in Democracy?
How local governance experiments in China preview the future of automated representation
A team of AI and sociology researchers at Renmin University of China (RUC) recently interviewed 92 residents of a Beijing urban community for two hours each, capturing migration stories, marriages, career turns, grievances about community issues. They transcribed these life-history interviews, then trained a small language model to answer governance surveys on behalf of those residents. The model gets it right about 50% of the time.1 They then deployed a live system where a residential committee can run the survey, generate a policy report, and revise the policy, without contacting any of the 92 people.2
The team just published their research as a preprint. It’s an exploratory social experiment, but given its direct test on actual local democracy in Chinese society, community-level decisions on parking, elevators, property management. It could also become the very first experiment of a future model for social participation.
A model of you
The technology to simulate a specific person inside a neural network was first built for the dead.
In 2016, Eugenia Kuyda, a software developer, gathered over 8,000 lines of text messages from her late friend Roman Mazurenko and fed them into a neural network. The result, “@Roman,” could carry on conversations in his voice — his turns of phrase, his idiosyncratic spelling. Friends texted it and found it unsettling in its accuracy.3 Kuyda then generalized the concept into Replika, a chatbot that learned to mimic living users’ personalities through ongoing conversation. By 2024, a full grief-tech industry had emerged: HereAfter, Eternos, and Chinese startups offering speaking avatars of the deceased for as little as ¥20. The premise that a person can be captured in model weights to speak in their absence had been accepted.
Powered by small models, these simulations of deceased people could still feel like general chatbots with slight personal tuning. However, by early 2025, when GPT-4.5 passed the Turing Test, the weaknesses of previous models were, in theory, resolved. An AI could now act like a human without being detected.
Soon after, Second Me, built by Mindverse and open-sourced in March 2025, became the first system designed to produce a digital twin of a living person. Users upload their data, notes, documents, audio, calendar, and feeds them into an automated pipeline that extracts entities, generates a biography, and creates synthetic question-answer pairs. With that data, the system fine-tunes a language model based on Qwen2.5-7B using LoRA (Low-Rank Adaptation, a technique that personalizes a large model by training only a small set of additional parameters). The result is a model whose weights have absorbed the user’s personality, patterns of speech, preferences, and reasoning habits. The model becomes a version of that person.4 Mindverse calls the result a “Context Provider.” It sits between the real person and other AI systems, enriching queries with personal context and representing the user in exchanges with other people’s digital twins.5 The architecture implies that a person can be sufficiently captured in an LLM to act as their proxy.
The delegate is already here
What enables AI to act in the real world is tool access. Equipped with tools, an AI can move beyond being a mere chatbot to actively take action and respond. Perplexity’s “Buy with Pro,” launched in November 2024, is a prime example of an AI performing specific tasks on behalf of humans.6 It lets users research and purchase products entirely within a conversation, while the agent handles all the backend execution. Users receive cited recommendations and spec sheets, completing checkout without ever visiting a retailer’s site. Despite Amazon’s legal challenge against Perplexity’s AI shopping tools, a U.S. appeals court recently ruled in favor of this model, concluding that the AI acts on behalf of the user rather than independently.7
Delegation has reached dating as well. Amata, a dating service with no swiping or browsing, launched in New York in September 2025. An AI matchmaker interviews users about their core values, lifestyle, and romantic goals, then proposes matches. When both sides agree to meet, each pays a $16 token, and the AI takes over the logistics, scheduling the evening, booking the restaurant. Direct messaging opens only two hours before the date; the user simply shows up.8 Justin McLeod, founder of Hinge, left his company in December 2025 to build Overtone along a similar model, backed by Match Group itself, betting that the swipe will be replaced by AI delegation.9
The bolder step to use agent-to-agent interaction, where the twins themselves do the talking has already been tried in dating but failed. Volar, founded by former Snap product manager Ben Chiang in late 2023, let users train a digital twin in five to ten minutes, voice clone included, then sent the twins on simulated first dates with each other while their owners watched and decided whether to step in as themselves. It shut down in September 2024, unable to raise further funding. The AI drifted could be part of the problem. A New York Times reporter found her twin had invented a trip to Japan she never took and declared a love of the Beatles she never had. The simulated dates came out unnatural and with endlessly recycling the same onboarding facts. And users never trusted what the other side’s AI, either.10
Beyond dating, Anthropic initiated an internal experiment in December 2025 called Project Deal, in which 69 employees each assigned a Claude AI agent to represent them in an office marketplace. The agents interviewed participants about their buying and selling preferences, set asking prices, made offers, countered bids, and closed transactions, all autonomously via Slack. The agents bargained over and finalized trades for real physical goods on behalf of their human principals.11
An agentic economy is taking root across daily life, from shopping and dating to complex negotiation. The trajectory follows a clear progression: from personalization, to human-to-agent delegation, to agent-to-agent interaction. Each domain follows the same pattern: the AI absorbs your preferences, acts on your behalf, and reports back. The question now is whether this pattern can enter the domain that sits at the center of human society: politics.
A party that listens to AI
If we view a political party as a company that produces policy, then corporate AI transformation methods can be seamlessly applied to politics.
Takahiro Anno, an AI engineer born in 1990, ran for Tokyo Governor in July 2024, placing fifth among 56 candidates with 150,000 votes. He subsequently founded Team Mirai (”Future”) in May 2025, winning 11 seats in the Lower House with 3.97 million votes (6.9% of valid ballots) in the February 2026 general election.12 During campaigns, Team Mirai deployed conversational AI avatars and interactive Q&A engines powered by Retrieval-Augmented Generation (RAG). These avatars answered thousands of voter questions 24/7 while keeping responses strictly grounded in the party’s manifesto. It functions essentially as automated customer service for constituents.
A good product comes from thorough market research. For Team Mirai, the core tool is an AI Interviewer that guides constituents through policy topics, probes their reasoning, challenges their assumptions, and generates structured summaries. Over 8,000 hours of constituent dialogue have been logged. In a single consultation regarding Diet reform, 36,000 messages were condensed into 13 key findings, a level of data compression no human campaign team could match.13
In this model, AI acts as a powerful assistant for a conventional political party, replacing traditional volunteer networks and massive campaign infrastructure. AI becomes a scaled ear: citizens speak, the system listens, and the party absorbs far more input than was previously possible. However, this is still merely how AI modernizes a political party within the existing framework. The AI Interviewer does not speak on behalf of anyone; it simply helps the party build a more “intimate” connection with the electorate.
What if AI spoke for you instead?
Team Mirai uses AI to carry voices in, the Renmin University paper creates a system that uses AI to generate voices.
The experiment has three parts. First, the input: trained fieldworkers interviewed each of 92 residents for two hours, producing a life-history narrative which includes personal background, education and career, family and marriage, community life and values. The design follows sociological life-course theory14, which posits that individual attitudes form progressively through accumulated life experiences.
Second, the target: each resident answered a 50-question survey co-designed with the residential committee around the nine issues of the community, which includes elevator funding, parking, property management, neighborhood disputes, volunteer work. Each question is single-choice with four to five scaled options: levels of support, satisfaction, frequency, or willingness.
Third, the results. Random chance on this survey instrument is 21.1%. LLMs operating without personal context score under 30%. Adding life-history narratives into the prompt raises accuracy by an average of 5.6 percentage points. Interestingly, incorporating a few of a resident’s prior answers and attitudes on other questions raises accuracy further. By combining life history with these reference answers, the strongest frontier models using pure prompting plateau near 49–50%. The team’s custom algorithm, Curriculum-LoRA, achieves 51.6% exact-match accuracy by fine-tuning a 7B parameter base model (Qwen2.5-7B-Instruct). Compared to traditional prompting, Curriculum-LoRA slashes costs by roughly 93 times relative to frontier models like GPT-4.1, making large-scale proxy simulation economically viable.
The researchers integrated this architecture into Onelink-Community, a public platform where officials author policy probes, execute simulations across virtual resident agents, and receive automated diagnostic reports on community acceptance and friction points. And it creates a closed loop from policy hypothesis to evaluation and back, without requiring direct contact with a single resident.
Joon Park at Stanford ran a similar study in 2024, with the same two-hour interview method, scaled to 1,052 Americans. The research reported 83% accuracy on predicting responses on the General Social Survey.15 83% seems quite higher than the RUC research, but Park’s questions were broad social attitudes, like abortion, religion, gun control. which people hold stably and correlate with life trajectory. And the 83% accuracy was also normalized against human self-consistency. People themselves only reproduce their own answers about 79.5% of the time when retested. So the actual raw exact-match accuracy is 65.7%. The RUC team 50% accuracy is generated upon predicting mundane local questions, without normalization.
As for the curriculum-LoRA method, it shares its core architecture with Second Me, which is also based on LoRA fine-tuning of the same Qwen2.5-7B base, with personal data absorbed directly into the model weights. However, Second Me was built as a general-purpose twin and has only been tested on its ability to recall life events. Thus, the RUC paper introduces a system that not only simulates a specific person, but also evaluates whether the simulation actually predicts their preferences.
How to increase the accuracy
Life-course sociology explains how attitudes form, but it is a weaker guide for predicting them individually. The paper shows that certain biography blocks alone score below the no-profile baseline, meaning a life story can be misleading. Moreover, incorporating a resident’s answers to adjacent questions creates a much larger leap in accuracy than their life story does, indicating that a person’s existing preferences predict their stances better than their background story.
Also, nobody holds a stable, well-formed position on low-salience issues such as local civic initiatives. People do not retrieve pre-existing opinions when surveyed; they construct them on the spot. Ask the same resident about parking satisfaction twice, two weeks apart, and their answers will not fully match. Joon Park at Stanford reported a 79.5% self-consistency rate on general survey questions. The RUC team never measured this self-consistency, but for mundane issues, it is reasonable to assume that self-consistency is even lower. If self-consistency were around 60%, the normalized accuracy of the RUC method would reach as high as 86.5%.
The fear of the crowd
The question of whether AI could serve as a political delegate lands differently in China than anywhere else, because China has spent a century building governance around a specific fear: that popular participation, unmanaged, produces chaos.
Published in 1924, Sun Yat-sen’s Three Stages of Revolution: 军政 (military rule), 训政 (political tutelage), 宪政 (constitutional government), assumed the population was not ready for democracy and required a period of party-led education.16 The tutelage stage had no deadline; the ruling party alone judged when society was “ready”. In Taiwan, martial law, a byproduct of this system, was only lifted in 1987.
After 1989 on the mainland, Deng Xiaoping stated that stability overrides everything, setting the trajectory of social governance all the way to today. Furthermore, a persistent argument regarding “population quality” (素质) maintains that the public’s capabilities are not yet sufficient for direct democratic participation. This concern is not only an elite talking point; it has shaped the architecture of Chinese governance for over a century.
And yet, China has always had local democracy. Village elections began with the 1987 Organic Law of Villagers’ Committees, giving hundreds of millions of rural residents their only experience of competitive balloting. Urban residents have had residents’ committees since 1954. Consultative democracy and the more recent framework of “whole-process people’s democracy” define legitimacy through consultation. This model of “controlled participation” was created to give citizens a voice while mitigating concerns over conflict.
AI delegation fits this architecture with precision. AI delegates are built on personal life histories, aligned periodically, and managed through official platforms. On the surface, this approach could provide more opportunities for people to participate in local governance without the crowd dynamics the state has long feared. The system could capture more voices, more often, at lower cost. It could even be described as a democratic upgrade: more residents consulted, more frequently.
Brave New World
At 7:15 AM, Lin sat at his dining table with a cup of soy milk, watching the morning light filter through the high-rises of District 7. His phone pushed a notification from MinGuan (Civil Voice), the official civic alignment application. It was time for his bi-weekly constituent preference calibration.
Lin unlocked the app and got an notification on the events of his recent life:
System: “We noticed your daughter registered for primary school enrollment next autumn, and your household electricity consumption during peak hours increased by 14%. Would you like to adjust your delegate’s priorities regarding sub-district education funding and municipal energy subsidies?”
Lin tapped Yes. He adjusted a few sliding toggles, prioritizing neighborhood school expansions over commercial park developments, and flagged his complaints over the construction noise along the Main Road. Very soon, Lin’s update was synced. Somewhere in a cloud datacenter, Delegate-Lin-2039, an AI delegate that trained on his decade digital footprint, voting history, and personal preferences was updated.
By the end of Friday, the Monthly Virtual Consultation Assembly for the Sub-district Urban Reform Act had concluded with a new proposal for its constituents to vote on.
Meanwhile, the physical District Congress building had long been converted into a public library. There were no more representatives in the local congress. The legislative process had been decentralized into a vast, continuous MinGuan network, where all of the district’s personal AI delegates convened.
Lin received a push notification on Saturday morning: Assembly Debrief & Feedback Log Ready. He tapped the screen to open the report. The app displayed a visual breakdown of how his delegate had performed during the virtual assembly:
Delegate Activity Report: Sub-district Urban Reform Act (Draft 4.2)
Alignment Score: 92.4% with your core interests.
Key Intervention: At 03:14:02 AM, during the automated debate on Zoning Scheme C, Delegate-Lin-2039 joined a coalition of 4,120 resident-agents representing young families.
Debate Record:
Municipal Planning Agent: Proposed allocating the vacant Northern Lot for a high-density commercial plaza to boost district revenue.
Delegate-Lin-2039: “Objection based on demographic trajectory data. Household profile #2039 and 68% of surrounding units require primary educational capacity within a 1.5 km radius by Q3 2040. Commercial zoning increases ambient noise pollution during study hours by an estimated 18 dB.”
Outcome: Delegate-Lin-2039 successfully negotiated a compromise clause: 60% of the lot reserved for an elementary annex, with ground-floor municipal retail to offset construction costs.
Lin scrolled through the condensed transcript, finding the exact words his AI delegate had spoken against the planning algorithms. His specific struggles, his daughter’s school commute, his sleep quality, had been articulated, argued, and defended. He hadn’t needed to make a speech, join a committee, or to to the streets. He was heard and empowered.
At the bottom of the screen, a red button illuminated: Vote. The AI had done the heavy lifting of negotiation and clause-drafting across thousands of conflicting neighborhood demands. Now it’s time for the final layer required by law: human vote.
Lin reviewed the synthesized ordinance, placed his thumb on the biometric scanner, and tapped Approve.
Across the sub-district, hundreds of thousands of residents were doing the exact same thing on their morning commutes, in their offices, and from their living rooms. No crowds had gathered. No tempers had flared. The noise of politics had been entirely absorbed by the machine, leaving behind only the quiet, seamless satisfaction of consent.
If my insights brought you a fresh perspective, please consider supporting me by buying me a coffee. Your generosity fuels my writing.
Chen, Xu, Yuanzi Li, Lei Wang, Nan Lu, Yang Wang, Anding Wang, Lei Shi, Xiaoxing Fu, and Ji-Rong Wen. "Benchmarking LLMs for Community Governance Simulation with Life-history Narratives." arXiv:2605.23783, Renmin University of China, June 2026. https://arxiv.org/abs/2605.23783
Onelink-Community, the deployed policy-simulation platform described in the paper. http://39.107.76.51:9882/
Newton, Casey. "Speak, Memory: When her best friend died, she rebuilt him using artificial intelligence." The Verge, October 6, 2016
Wei, Jiale, Xiang Ying, Tao Gao, Felix Tao, and Jingbo Shang. "AI-native Memory 2.0: Second Me." arXiv:2503.08102, Mindverse.ai, March 2025. https://arxiv.org/abs/2503.08102
Second Me open-source repository. Mindverse, released March 2025. https://github.com/mindverse/Second-Me
Perplexity. "Shop like a Pro." Company announcement of Buy with Pro, November 2024. https://www.perplexity.ai/hub/blog/shop-like-a-pro
Bloomberg Law. "Perplexity Overturns Amazon Ban on AI Shopping Bot on Appeal." August 4, 2026. (Ninth Circuit vacates the injunction; users, not Perplexity, "access" Amazon under the CFAA.) https://news.bloomberglaw.com/us-law-week/perplexity-overturns-amazon-ban-on-ai-shopping-bot-on-appeal
Business Insider. "A new AI matchmaker will set you up and plan the first date — skipping the swiping and DMs." September 29, 2025. (Amata's NYC launch; the $16 token; the two-hour DM window; cancellation rules.) https://www.businessinsider.com/new-ai-matchmaking-dating-app-amata-launches-us-new-york-2025-9
TechCrunch. "The founder of Hinge raised $18M to build a new AI dating service, Overtone." July 14, 2026. https://techcrunch.com/2026/07/14/the-founder-of-hinge-raised-18m-to-build-a-new-ai-dating-service-overtone/
Tan, Eli. "Are A.I. Clones the Future of Dating? I Tried Them for Myself." The New York Times, November 14, 2024. (Volar's clone-to-clone dates; the clone's invented Japan trip and Beatles enthusiasm; Volar's September 2024 shutdown.)
Anthropic. "Project Deal: our Claude-run marketplace experiment." April 2026. (69 employees, 186 deals, over $4,000 transacted via Claude agents in December 2025.) https://www.anthropic.com/features/project-deal
Sanders, Nathan E., and Bruce Schneier. "Rewiring Democracy Now." Schneier on Security, January 12, 2026. (Team Mirai's AI Interviewer, the Diet-reform consultation, Digital Democracy 2030.) https://www.schneier.com/essays/archives/2026/01/rewiring-democracy-now.html
Sanders, Nathan E., and Bruce Schneier. "Japan's Team Mirai Uses Tech to Bolster Democracy, Not Undermine It." Tech Policy Press, March 19, 2026. (8,000+ hours of AI Interviewer engagement; the February 2026 election result.) https://www.techpolicy.press/japans-team-mirai-uses-tech-to-bolster-democracy-not-undermine-it/
Elder, Glen H., Jr. "The Life Course as Developmental Theory." Child Development 69, no. 1 (1998): 1–12. (The sociological life-course framework underlying the RUC interview design.)
Park, Joon Sung, Carolyn Q. Zou, Jonne Kamphorst, Niles Egan, Aaron Shaw, Benjamin Mako Hill, Carrie Cai, Meredith Ringel Morris, Percy Liang, Robb Willer, and Michael S. Bernstein. "LLM Agents Grounded in Self-Reports Enable General-Purpose Simulation of Individuals" (originally circulated as "Generative Agent Simulations of 1,000 People"). arXiv:2411.10109, Stanford University, November 2024.
Sun Yat-sen. Fundamentals of National Reconstruction (建国大纲), 1924. (The three stages: 军政, 训政, 宪政.)