ConnieChan.ai
ConnieChan.ai was shut down on August 7, 2026, four days after launching as a satirical political chatbot.
“In our view, spending twenty grand on billboards to direct voters to an unhinged chatbot that immediately insulted its subject's citizenship is an all-time masterclass in political self-sabotage.”
Our takeConnieChan.ai was a satirical political chatbot launched in August 2026 by Scott Wiener's congressional campaign to parody opponent Connie Chan, which was taken offline within days after generating racially insensitive outputs and sparking fierce backlash.
- Hype · our take
- 3/5 promised → 1/5 delivered
- Lifespan
- 4 days
- Timing · our take
- right time,
wrong execution
- Dates
- announced · shipped · discontinued
- Where
- US · Startup · Internal
The Pitch
ConnieChan.ai was a satirical web application launched by California State Senator Scott Wiener during his campaign for the United States Congress E4E4E4. Built to caricature San Francisco Supervisor Connie Chan in the race to succeed Nancy Pelosi, the system was presented as political satire E4. On its homepage, the tool billed itself as Connie, the world's first AI trained to say no to anything you ask E4. The application was powered by Anthropic's Claude and operated as a free conversational interface with no paid tier E4. Visitors interacted with the chatbot via a text input box on the web where they could submit questions about local governance or select pre-set discussion topics E4. Recommended prompts were engineered to frame Chan as an obstructionist, directing inquiries toward themes such as blocking housing, derailing public transit, failing education, and stalling the city's economic recovery E4. The chatbot was designed to mimic Chan by rejecting every user request and proposal E4. Rather than offering neutral election information, the software sought to satirize her municipal record on housing development and transit projects to persuade voters in San Francisco's 11th Congressional District E4.
What Happened
The chatbot went live online on August 3, 2026, accompanied by an outdoor marketing campaign E4. Wiener's campaign spent approximately $20,000 on two physical billboards in San Francisco promoting the tool under the slogan "Let’s do nothing together" E4. Traffic surged immediately upon release, with users submitting more than 5,000 questions to the system during its initial 27 hours online E4. The reaction from political opponents was swift and intensely critical E4. Chan's campaign condemned the release as "tired and trite slop" and accused Wiener of being bought and paid for by the technology industry E4. Chan's campaign and political allies further denounced the synthetic avatar as racist and offensive to women E4. In response to early criticism, Wiener's campaign initially defended the project, stating that no lies were told in the making of the chatbot and maintaining that it merely packaged Chan's voting record for voters E4. However, serious operational issues quickly emerged from the bot's unconstrained model responses E4. When users questioned the synthetic persona about her accent, the model generated jokes claiming its accent was so thick it could barely understand itself E4. When asked whether Chan was a United States citizen, the system offered evasive answers suggesting the topic required further study and a public hearing E4. Scrutiny also turned to data handling, as the site's privacy terms allowed staff to retain user conversations for up to 90 days E4. Representative Nancy Pelosi issued a public statement condemning the tool as a severe abuse of technology that relied on gendered and racial stereotypes E4. Facing escalating condemnation, Wiener pulled the plug on the chatbot and agreed to remove the promotional billboards E4. By August 7, 2026, just five days after launch, the campaign took ConnieChan.ai offline permanently E4E4E4E4. Wiener acknowledged that the political experiment had "missed the mark" and apologized to those offended, stating that the project had distracted from core policy debates E4E4.
Why It FailedOur take
In our view, ConnieChan.ai failed primarily because of public backlash and the severe reputational hazards inherent in political impersonation deepfakes. Scott Wiener's campaign sought to package opposition research into an interactive parody, but delegating political messaging to an unconstrained foundation model created catastrophic safety risks. The bot quickly produced culturally insensitive commentary about accents and citizenship, transforming a targeted campaign gimmick into an ethics scandal involving accusations of racism and sexism. In our view, secondary failures stemmed from bias scandals and uncontrolled prompt boundaries, which alienated key party allies including Nancy Pelosi. When an AI stunt creates liabilities faster than it conveys policy arguments, pulling the product offline is the only politically viable recourse.
Lessons and What Would Have Worked
- In our view, political campaigns should avoid deploying synthetic personas of living opponents, as algorithmic responses inevitably escape script boundaries and trigger severe backlash.
- Founders and campaign teams wrapping third-party language models must enforce rigid refusal boundaries on sensitive demographic topics like citizenship and national origin before public release.
- When a promotional AI artifact generates negative headlines that overshadow the underlying product or campaign message, immediate shutdown is more effective than attempting iterative guardrail patches in production.
Timeline
- Scott Wiener's campaign launches ConnieChan.ai alongside $20,000 in promotional billboards.E4
- Chan's campaign condemns the chatbot as trite slop and Wiener's team defends the parody.E4
- Criticism mounts over racial insensitivity and offensive responses generated by the bot.E4
- Wiener announces the chatbot is being taken offline and apologizes after fierce backlash.E4
What happened to ConnieChan.ai?
- What was ConnieChan.ai?
- ConnieChan.ai was a satirical political chatbot launched in August 2026 by Scott Wiener's congressional campaign to parody opponent Connie Chan, which was taken offline within days after generating racially insensitive outputs and sparking fierce backlash.
- What happened to ConnieChan.ai?
- The chatbot went live online on August 3, 2026, accompanied by an outdoor marketing campaign .
- Why did ConnieChan.ai fail?
- In our view, ConnieChan.ai failed primarily because of public backlash and the severe reputational hazards inherent in political impersonation deepfakes.
Receipts
- E4Scott Wiener launches Connie Chan chatbot - AOL
Supporting quote
SAN FRANCISCO ( KRON) — As the race to succeed Nancy Pelosi in Congress heats up, State Sen. Scott Wiener has launched an AI chatbot parody of his opponent, SF Supervisor Connie Chan. The chatbot, ConnieChan.ai, describes itself as “Connie, the world’s first AI trained to say no to anything you ask.”
- E4Nancy Pelosi’s pick for Congress skewered as ‘say no to anything’ pol by AI chatbot - AOL
Supporting quote
AI is supposed to solve our problems — and a cheeky new chatbot says Connie Chan is one of them. California state Sen. Scott Wiener launched a parody artificial intelligence chatbot this week mocking his rival Chan, a San Francisco supervisor, escalating the bitter fight between the two Democrats seeking to succeed longtime Rep. Nancy Pelosi in Congress.
- E4Scott Wiener pulls AI chatbot parody of rival Connie Chan after backlash
Supporting quote
California state Sen. Scott Wiener is pulling the plug on a [controversial AI chatbot](https://nypost.com/2026/08/08/us-news/nancy-pelosi-rages-at-sexist-and-racist-ai-chatbot-mocking-her-would-be-successor/) and the billboards promoting it after his campaign faced a furious backlash over the parody targeting his [congressional rival Connie Chan](https://nypost.com/2026/08/05/us-news/nancy-pelosis-pick-for-congress-skewered-by-ai-chatbot/). Wiener, who is running to replace former [House Speaker Nancy Pelosi](https://nypost.com/2026/08/08/us-news/nancy-pelosi-rages-at-sexist-and-racist-ai-chatbot-mocking-her-would-be-successor/) in San Francisco’s 11th Congressional District, announced Friday his campaign would remove ConnieChan.ai, along with the ads directing voters to it. The chatbot was created as a campaign attack on Chan, portraying what Wiener said was her record on housing, public transportation and other issues.
- E4彭博罵AI冒犯性下架 - ss future
Supporting quote
一個由尋求接替前美國眾議院議長南希・佩洛西（Nancy Pelosi）議席的候選人推出、帶有諷刺意味的人工智能聊天機械人，因佩洛西炮轟其內容「性別歧視」而在周五被下架。 加州州參議員 Scott Wiener 本周推出這個聊天機械人，目的係嘲諷同屬民主黨、兼任市議員的 Connie Chan。Chan 早前獲佩洛西背書。Wiener 同 Chan 將於 11 月爭奪佩洛西留下嘅席位。 網站 ConnieChan.ai 自稱係「全球首個受訓去拒絕你任何要求的 AI」，明顯係針對 Chan 俾批評者形容為進步派 NIMBY（鄰避主義者）嘅形象，暗示佢過往多次反對房屋發展方案。 Chan 陣營當然唔會欣賞 Wiener 呢種極不尋常嘅競選手法，而佩洛西亦隨即出聲，公開表達對呢個做法嘅憤怒。 佩洛西周四發表長篇聲明，並獲多名女性民選官員支持。她指出：「女性幾代人以來一直努力，先至可以喺權力殿堂入面有自己嘅聲音。我哋唔應該而家仲要同人工智能鬥，去阻止有人假借我哋嘅名義講嘢。」 聲明批評 Wiener 利用 AI 模仿女性聲音，形容呢做法係「對科技嘅嚴重濫用，亦反映出一種令人不安、想抹去女性主體性嘅態度」。聲明更指，由於 Chan 係有色人種女性，相關傷害會同「種族刻板印象」疊加，令問題更加嚴重。
- E4Scott Wiener Takes Down Ill-Fated AI Chatbot
Supporting quote
State Senator Scott Wiener pulled down a controversial artificial intelligence chatbot after backlash from Nancy Pelosi and others. The campaign had briefly rolled out ConnieChan.ai, a satirical bot meant to mimic San Francisco Board of Supervisor Connie Chan and answer questions about her record. The site's homepage called it "the world's first AI trained to say no to anything you ask," and the campaign promoted it with $20,000 worth of billboards in San Francisco. Chan's team condemned it, calling it an example of Wiener being "bought and paid for by Big AI." Wiener had defended the bot as "clearly a parody" and said Pelosi was making "baseless" claims. He also referred to California Senate Bill 1047, a pro-regulation AI safety bill that would have made companies legally liable for harms posed by their AI. By Friday afternoon, five days after the site launched, the campaign agreed to take it down. Wiener said, "Today, we are taking down the chatbot, and the billboards that promoted it will come down soon," and apologized after saying the debate had shifted away from housing and affordable healthcare.
- E4A Politician Built an AI Version of His Rival. It Didn’t Go as Planned. - AI Updates
Supporting quote
In early August, California State Senator Scott Wiener, a Democrat running for Congress, launched ConnieChan.ai, a website featuring a chatbot designed to imitate San Francisco Supervisor Connie Chan. Powered by Anthropic’s Claude, the bot was not intended to provide neutral information about the race. It was built as political satire, portraying Chan as a politician predisposed to oppose housing, development, transit projects and other initiatives associated with Wiener’s political agenda.
Sentiment
At launch
No launch coverage quote recorded.
At death
No death coverage quote recorded.
No sentiment quotes recorded.
Last updated · Phil · How we source entries
Changelog
- — Published ConnieChan.ai.