Boundaries and Responsible Use of AI (Disclaimer)
AI AGENT Role & Boundaries
What AI AGENT does:
Analyzes your profile
- Goals: capital preservation / passive income / growth
- Risk tolerance: conservative / balanced / opportunistic
- Time horizon: short-term / long-term
- Preferences: chains, assets, lock-up constraints
- Experience level: beginner / intermediate / advanced
Filters protocols
- Allow-list: Only shows audited, vetted protocols
- Risk scoring: Based on TVL, audits, time alive, incidents
- Matching: Aligns with your declared profile
- Excludes: Quarantined/high-risk options automatically
Shows options
- Option A: "Best match" for your profile
- Option B: Alternative (different risk/reward balance)
- Option C: Alternative (different chains/protocols)
- Each with: Expected APY, costs, risks, steps
Explains why ("Reason-codes")
- "Recommended because: [your goal] + [low risk] + [high liquidity]"
- "Alternative because: [higher APY] but [lock-up period]"
- "Avoid because: [recent incident] / [low liquidity] / [high cost]"
Monitors & suggests adjustments
- "APY dropped 2% → Consider migration?"
- "New protocol matches your profile → Review?"
- "Gas prices low → Good time to rebalance?"
- YOU approve any changes (AI doesn't execute)
What AI AGENT does not do:
Does not provide financial advice
- AI suggestions are INFORMATIONAL, not investment advice
- You are responsible for your own decisions
Does not guarantee returns
- All yields are variable and protocol-dependent
- AI cannot predict future market conditions
Does not execute transactions
- AI prepares transaction intents
- YOU must sign every transaction in your wallet
Does not have access to your funds
- AI cannot move your money
- Non-custodial model maintained
Does not replace due diligence
- Always read protocol docs, understand risks
- AI is a tool to help, not a substitute for research
Why "Informational Only" Matters
Legal & Regulatory
| If AI Were "Advisory" | OROKAI's "Informational" Approach |
|---|---|
| Would require investment advisor license (SEC, FCA, MAS...) | Software providing data & tools |
| Fiduciary duty (must act in client's best interest) | User makes own decisions |
| Liability for losses ("bad advice") | Disclosures: variable returns, user responsibility |
| Mandatory suitability assessments | Optional profile for personalization |
| Ongoing monitoring obligation | Monitoring as a feature, not obligation |
| Detailed compliance reporting | Standard software ToS & Privacy Policy |
Regulatory Precedent:
- Robinhood "Learn": Educational content, not advice → No RIA license needed
- CoinGecko alerts: "BTC dropped 5%" notification → Informational, not advice
- TradingView indicators: Technical analysis tools → User interprets, not advice
- OROKAI AI: Protocol suggestions with reasoning → User decides, not advice
Scope and Limitations
No investment advice
Recommendations are informational in nature. Decisions and authorizations remain with the user.
No return/price guarantees
Returns and rates are variable and depend on protocols and the market; the agent does not guarantee best price or execution.
Non-custodial
Agent does not have access to keys and does not execute transactions; operates in a prepare-and-explain model.
Security Controls ("Guardrails")
Protocol allowlist
Allow-list of protocols/bridges/DEX, exposure limits per protocol and chain.
Pre-signature verifications
List of steps, costs (gas/slippage/fees), required smart contract permissions.
Anomaly detection
Risk signals (liquidation, liquidity, incidents) → pause and warn.
Permission revocation
Quick links to revoke/limit approvals.
Compliance
Geo-restrictions, sanctions, partner rules (on/off-ramp, cards) on their side; we enforce restrictions in UI.
AI Agent OROKAI provides information and proposals aligned with the user's declared profile. This is not investment advice. All operations require the user's conscious signature in their wallet. Returns are variable and not guaranteed.