🤖

AI Services Terms Reviews

Your prompts, their training data? AI services raise novel questions about content ownership, output rights, and how your inputs may train future models. I analyze the terms that define your rights in the AI era.

6
Services Reviewed
55
Avg Score
Anthropic
Best in Category

Service Comparison

Side-by-side comparison of consumer fairness scores across major AI services.

Service Score Grade Key Gotcha
Anthropic (Claude)
Best in Category
72 B Clearest training data policies, transparent terms View Review →
Perplexity AI
62 B- Citations don't guarantee accuracy View Review →
OpenAI (ChatGPT)
55 C+ Training data opt-out complexity, frequent policy changes View Review →
GitHub Copilot
52 C Code ownership ambiguity, license detection limits View Review →
Google Gemini
48 C Google ecosystem data integration View Review →
Midjourney
42 C- Public visibility default, Discord-based support View Review →

🔎 Key Findings Across All Services

Output Ownership Varies

Most services grant you rights to outputs, but with limitations. Commercial use, attribution requirements, and ownership of AI-generated content remain legally unsettled.

Training Data Questions

Your inputs may train future models unless you opt out. Opt-out processes vary in complexity and completeness across providers.

Content Restrictions

All services impose content policies that can restrict output or terminate access. The boundaries of "acceptable use" are often subjectively enforced.

No Accuracy Guarantees

Every service disclaims responsibility for AI "hallucinations" and inaccurate outputs. You bear full liability for how you use AI-generated content.

Service Reviews

Click through for complete analysis of each service's terms.

Anthropic (Claude)

72/100 (B)
  • Clearest training data policies
  • Constitutional AI safety focus
  • Transparent usage limits
Full Review →

Perplexity AI

62/100 (B-)
  • Citations don't guarantee accuracy
  • Uses multiple third-party AI models
  • Clear pricing structure
Full Review →

OpenAI (ChatGPT)

55/100 (C+)
  • Training opt-out complexity
  • Output ownership granted to user
  • Frequent policy changes
Full Review →

GitHub Copilot

52/100 (C)
  • Code suggestions may match existing code
  • License detection limitations
  • Indemnification for Business only
Full Review →

Google Gemini

48/100 (C)
  • Deep Google ecosystem integration
  • Human review of conversations
  • Multi-layered terms complexity
Full Review →

Midjourney

42/100 (C-)
  • Images public by default
  • Commercial use restrictions
  • Discord-based, limited support
Full Review →

AI Service Questions?

Questions about AI output ownership, training data rights, or enterprise licensing? I can help you understand the legal landscape of AI services.

Scan Your ToS

Start with a written review

Need help understanding your rights or resolving a platform dispute?

$240 Written Attorney Consultation
Email Sergei with documents
Live interactive demo

Try an AI governance contract workroom

After scoring these AI services' terms for fairness, I turn the same clauses into an editable AI governance contract workroom. Set autonomy to fully autonomous, or turn vendor training on your data on, and watch the room flag the boundary in real time: live preview with surgical yellow highlighting, click-any-clause comments, and track-changes style suggestions.

Open the live demo workroom How I build these for firms
Fictional demo data. Built by Sergei Tokmakov, Esq., California attorney and AI engineer.
How this is scored. I built the methodology; an automated system scores each company's published terms against it. Treat these scores and flags as a consistent, opinion-based read, not a guarantee or legal advice. Terms change often, so check the company's current policy before relying on this.