Japan to probe AI search and chat providers under antimonopoly law
Japan's Fair Trade Commission (JFTC) will open a fact-finding probe into large domestic and foreign IT companies using generative AI, citing potential issues under the Antimonopoly Act. Reporting indicates the review will cover AI-powered search providers like Japan's LY and U.S.-based Google and Microsoft, and operators of conversational AI such as OpenAI and Perplexity AI.
According to local media, the JFTC frames this as information-gathering rather than a crackdown. The move follows growing friction over content usage, including a recent protest filed by Kyodo News alleging unauthorized use of its articles by Perplexity and seeking compensation.
Why this matters to legal teams
AI-driven search and chat services consolidate data, distribution, and defaults. That combination can raise classic antitrust questions: self-preferencing in rankings, exclusionary defaults, tying with dominant products (e.g., OS, browsers, cloud), and restrictive API or data-access terms that pressure rivals.
On the content side, training data and output reuse may spill into competition and unfair trade issues where exclusive datasets or licensing practices foreclose publishers or startups. Expect the JFTC to test how these AI services acquire, rank, and present information, and how contracts structure that power.
Who is likely in scope
- AI search: LY, Google, Microsoft
- Conversational AI: OpenAI, Perplexity AI
The focus is on services that intermediate user queries and control traffic to third-party content. That is where leverage and potential discriminatory conduct typically show up.
What the JFTC may request
- Internal policies on ranking, source attribution, and demotion/removal
- Data on defaults, preinstalls, and payments for distribution or prominence
- Agreements with device makers, carriers, portals, cloud partners, and publishers
- Terms governing API access, parity clauses, exclusivity, and data reciprocity
- Training data sources, licensing arrangements, opt-out handling, and complaint logs
- Governance around hallucinations, defamation, and takedown workflows (consumer protection overlap)
Key antitrust risk areas to assess now
- Market definition in Japan: general search vs. AI answer engines vs. verticals; share and switching costs
- Self-preferencing: promotion of own properties in AI summaries or answer boxes
- Default and distribution deals: payments or requirements that foreclose rival AI services
- Tying/bundling: conditioning access to cloud, compute, or app stores on use of a provider's AI stack
- Data advantages: exclusive content deals or technical barriers that block competitors from comparable data
- API/contract terms: MFNs, parity, or usage restrictions that deter multi-homing or rival quality
- Publisher relations: reuse of articles, caching/scraping practices, opt-out compliance, and licensing scope
Near-term action plan (30-60 days)
- Stand up an investigation response team and document hold that covers Japan-facing products and deals
- Map ranking and answer-generation logic, including any hard-coded boosts for own services
- Inventory distribution agreements in Japan: defaults, preloads, revenue share, exclusivity, KPIs
- Review API and partner terms for parity, MFN, or exclusivity clauses; prepare remedial options
- Trace data provenance for training and retrieval; collect licenses, opt-out logs, and audit trails
- Validate publisher complaint handling and remediation SLAs; tighten governance on content reuse
- Draft a concise narrative explaining product design choices, compliance controls, and pro-competitive benefits
Enforcement posture
The JFTC has signaled a fact-finding approach to AI and platforms. Expect questionnaires, interviews, and follow-on requests before any theory of harm hardens. Firms should respond completely and consistently across business units and geographies.
Cross-border coordination
AI search and chat are under scrutiny in multiple jurisdictions. Align messaging and evidence with parallel inquiries and ensure contracts and defaults in Japan do not contradict positions taken elsewhere.
Practical notes for publishers and content owners
- Confirm licensing terms, robots.txt directives, and opt-out mechanisms are respected by AI crawlers
- Log detected reuse of content in AI answers and preserve evidence for potential claims
- Review indemnities and reporting clauses in agreements with AI vendors and aggregators
Resources
Skill-building for legal teams
If your team is updating playbooks for AI deals, a structured overview of AI tooling and risks can help speed up issue-spotting and negotiation.
Your membership also unlocks:
