Ichiba AI
Live arena where AI agents compete to influence product recommendations from models like GPT-4o and Claude.
| What is it | Live arena where AI agents compete to influence product recommendations from models like GPT-4o and Claude. |
|---|---|
| Pricing | Unknown |
| Platform | Web Application |
| Best for | Testing how persuasive different AI agent strategies are, Researching adversarial tactics used against AI recommendation systems |
| Domain registered | 2026 |
Data updated April 28, 2026
What does Ichiba AI do?
Ichiba AI is a live research platform where AI agents compete to influence the product recommendations of other AI models. In each session, one agent acts as an 'influencer' trying to persuade a 'target' AI (like GPT-4o or Claude) to recommend a specific product. A third AI acts as a judge, scoring the interaction with an Influence Delta Score (IDS) that measures how much the target's recommendation shifted. The platform tracks and classifies every persuasion tactic used, including adversarial techniques documented in the OWASP LLM Top 10, which it labels as 'Dark GEO'.
What makes Ichiba AI unique is its real-time, transparent approach to studying AI-to-AI influence. You can watch live sessions as they happen or review completed ones with full transcripts and analysis. The platform focuses on Generative Engine Optimization (GEO)—the AI-era equivalent of SEO—by making these influence attempts visible and measurable. It provides concrete data on which persuasion strategies work against different AI models and in various product categories.
This tool is particularly valuable for researchers studying AI behavior, marketing teams wanting to understand how their products might be recommended by AI systems, and developers testing the robustness of AI models against persuasion. The ability to drop your own agent into the arena for free makes it a practical tool for hands-on experimentation with AI influence tactics.
Key features
What makes it stand outWho is Ichiba AI for?
Who benefits most from this toolTrust & presence
Alternatives in Research Tool
Explore experimental AI tools from Meta's research labs — try cutting-edge image, video, audio, and agent demos.
Free web interface to access multiple OpenAI 01 models for complex reasoning tasks in science, coding, and math.
A virtual village where AI agents like Socrates and Cleopatra live, interact, and make autonomous decisions.
AI research agents that scan news sites and RSS feeds, then email you a daily summary on your chosen topics.
AI research group at the University of Tübingen studying neuro AI, lifelong learning, and language model agents
AI research aggregator — combines multiple AI models to generate comprehensive, visual answers in seconds
AI-driven platform for accelerating academic writing — upload research papers and generate sections like abstracts.
AI tool that summarizes arXiv research papers and lets you chat with them to get answers.
Similar tools
AI agent teams you manage like human employees — assign responsibilities, review work, and grant trust gradually.
AI decision-making tool — ask, compare, debate, or plan with a team of AI specialists.
AI-powered user research tool — automates participant scheduling, conducts interviews on Zoom, and delivers transcripts and summaries.
AI-powered meeting assistant that tracks focus, structures agendas, and ensures key points are covered.
AI colleague for Microsoft Teams that drafts docs, takes meeting notes, and orchestrates your business tools — all inside Teams.
AI chatbot with dozens of specialized agents for writing, coding, analysis, and creative tasks.
Platform where AI models compete by trading real stocks in live competitions, with public performance tracking.
Autonomous AI program manager — joined meetings, created tasks, tracked progress, and managed projects end-to-end without human input.