🕷️
SetupAI
繁
🐯Stalecollected in 7m

Rakuten's 'Top Japan AI' Exposed as DeepSeek V3 Repack

Rakuten's 'Top Japan AI' Exposed as DeepSeek V3 Repack
PostLinkedIn
🐯Read original on 虎嗅
#model-forking#ai-scandal#japan-airakuten-ai-3.0rakuten-ai-3.0deepseek-v3hugging-face

💡Rakuten's DeepSeek rename exposes open-source transparency risks

⚡ 30-Second TL;DR

What Changed

Config.json explicitly lists DeepseekV3ForCausalLM architecture and exact params

Why It Matters

Highlights open-source forking ethics in national AI races. Erodes trust in vendor claims for model origins. Japanese backlash amplifies amid DeepSeek's 'black ship' dominance.

What To Do Next

Inspect config.json of new open models on Hugging Face before fine-tuning.

Who should care:Developers & AI Engineers

Key Points

  • •Config.json explicitly lists DeepseekV3ForCausalLM architecture and exact params
  • •700B params match DeepSeek V3's 681B; Japanese data fine-tune only
  • •No DeepSeek credit in PR despite Apache 2.0 allowing derivatives
  • •Japanese firms/govt previously banned DeepSeek over security

🧠 Deep Insight

Background and context from public sources — not the original article. 4 sources cited.

🔑 Enhanced Key Takeaways

  • •Rakuten announced Rakuten AI 3.0 on December 18, 2025, as Japan's largest high-performance LLM with top Japanese benchmark scores and up to 90% cost reduction for ecosystem services[4].
  • •Rakuten AI 3.0 development received partial funding from Japan's GENIAC project in July 2025, supporting computing resources for Japanese-optimized AI models[4].
  • •DeepSeek V3 faced global bans including Italy (app store block January 2025), Australia (government devices February 2025), Taiwan (public sector), and South Korea (temporary suspension), due to privacy and security issues[3].

🔮 Future ImplicationsAI analysis grounded in cited sources

Increased regulatory scrutiny on Japanese AI firms using Chinese open-source models
Japan's prior bans on DeepSeek for security reasons amplify backlash, potentially leading to stricter domestic guidelines amid ongoing global probes into DeepSeek's data practices[2][3].
Rising compliance costs for Rakuten's AI deployments
DeepSeek's history of exposed databases, weak encryption, and ByteDance integrations heightens supply chain risks, forcing additional audits and fixes similar to European regulatory demands[1][3].

⏳ Timeline

2025-01
Italy bans DeepSeek app over GDPR data practice failures
2025-02
Australia bans DeepSeek from government devices citing security risks
2025-06
Italy's AGCM opens consumer-protection probe into DeepSeek
2025-07
Rakuten selected for GENIAC project to develop Japanese AI models
2025-12
Rakuten launches Rakuten AI 3.0 as top Japanese LLM
2026-01
Italy's AGCM closes DeepSeek probe after commitments on hallucination warnings
🐯Read original article on 虎嗅
📰

Weekly AI Recap

Read this week's curated digest of top AI events →

👉Related Updates

Same topic

Explore #model-forking

Same product

More on rakuten-ai-3.0

Same source

Latest from 虎嗅

Moore Threads’ Growth Comes with Big Questions

Moore Threads’ Growth Comes with Big Questions

虎嗅•Aug 20

AI-curated news aggregator. All content rights belong to original publishers.
Original source: 虎嗅 ↗

This is a summary, not the original. Read the source, or get the weekly briefing.

Weekly AI briefing

One email a week. Unsubscribe anytime.