Leonardo.AI Enterprise Risk Assessment | AI Tool Risk

Back to AI Tools

Leonardo.AI

Risk Overview

Moderate Risk

4.4

/10

Image Generation

Top Risks

Binding arbitration with class-action waiver for non-Australian users (Legal Terms, score 7) creates friction for dispute resolution and limits collective remedies; disputes referred to Sydney-based ACICA arbitration under Australian law, with separate Irish law handling for EU data. Absence of publicly confirmed SOC 2 Type II or ISO 27001 certifications (Security Posture, score 5) despite enterprise DPA availability raises questions about verifiable security standards for customer data.

Mitigations

For legal protection, negotiate enterprise MSA/DPA terms explicitly addressing dispute resolution preferences before adoption; document any security requirements in procurement. Use Leonardo's paid private tier (Essential $12+/month) to ensure content is excluded from AI training and user retains full IP ownership; enable private-by-default mode and audit public content sharing settings regularly. Monitor the availability of SOC 2/ISO 27001 certifications via vendor trust portal and request audit summaries as part of enterprise procurement—current absence does not preclude use for non-sensitive workloads but should inform data classification decisions.

Get started

Is your team using

Leonardo.AI

safely?

Book a Demo Free Assessment

Highest Risk Categories

The highest-scoring risk categories identified in this assessment.

Legal Terms

Disputes for non-Australian users referred to arbitration administered by Australian Centre for International Commercial Arbitration (ACICA), conducted in Sydney under ACICA Arbitration Rules. EU/UK data disputes handled under Irish law courts. Class action waiver implied by individual arbitration requirement. Governing law is complex (Australian for most, Irish for EU data). Liability limitations apply. Less aggressive than US-style arbitration clauses but still binding arbitration required.

/10

Security Posture

No major publicly disclosed data breaches found. Security profile reviewed by Nudge Security showing basic security posture. Data Processing Addendum available indicating some enterprise security considerations. No publicly confirmed SOC 2 Type II or ISO 27001 certifications found in research. Standard cloud security practices assumed (likely uses AWS/GCP infrastructure). No bug bounty program identified in research.

/10

Training Use

Private content for paid subscribers is explicitly NOT used for AI model training without express written consent. Public content shared on the platform can be used by Leonardo and the community. Free tier content in public mode is more exposed. The distinction between public and private content is clear and contractually protected for paid users. Users must actively choose to make content public.

/10