It's been another incredible week at Perle. In addition to the close of our wildly successful Perle Human CAPTCHA campaign, we hit 300K on twitter. We could not do this without each and every single one of you. It was so amazing to meet you IRL in Denver, HK, and Korea. But we are just getting started. For now, we have tasks. And more exciting announcements to share very soon.

‍

Contracts aren't written in isolation. Clauses interact — and sometimes they conflict. A confidentiality provision might clash with a disclosure requirement. A liability cap might contradict an indemnity carve-out. These inconsistencies create real downstream risk.

‍

‍

AI systems analyzing legal documents need to perform cross-provision reasoning, not just read clauses in isolation. That requires human judgment to identify where language, timelines, and obligations don't align.

‍

The Multi-Issue Clause Conflict Detection Quest is now live on Perle Labs! Visit app.perle.xyz to start earning points.

‍

What This Task Is

The Multi-Issue Clause Conflict Detection Quest is a set of 10 legal analysis tasks. Each prompt presents two contract clauses from the same agreement. The clauses may work together harmoniously — or they may conflict in subtle or significant ways.

‍

Your job is to determine whether a conflict exists and, if so, identify the type. These classifications train AI systems to detect internal inconsistencies in legal agreements and identify ambiguity before it creates contractual risk.

‍

What You'll Do

‍

Step 1: Read both contract clauses carefully.

Step 2: Compare language, defined terms, timelines, obligations, exceptions, and carve-outs.

Step 3: Determine whether a conflict exists between the two clauses.

Step 4: If a conflict exists, identify the type and submit. If no conflict exists, indicate that the clauses are consistent.

Some clauses appear conflicting at first glance but are reconciled through "notwithstanding" language or explicit carve-outs. Read carefully.

‍

Conflict Types

• Contradictory obligations: Clauses require opposite actions
• Scope overlap creating ambiguity: Unclear which clause governs
• Definitional inconsistency: Same term defined differently
• Jurisdictional inconsistency: Conflicting governing law or venue
• Timeline conflict: Incompatible deadlines or durations
• Cap vs. uncapped exposure conflict: Liability limits contradict indemnity scope
• Carve-out tension: Exception in one clause undermines another
• Other structural inconsistency: Misalignment not covered above

‍

Why It Matters

By contributing to this task, you are:

• Teaching AI to perform cross-provision reasoning in legal documents
• Building enterprise-grade datasets for contract review and risk analysis systems
• Improving how AI detects ambiguity and internal inconsistencies before they create problems
• Contributing to the sovereign data layer that legal and enterprise institutions require

‍

This is expert-level work that AI cannot learn alone — the human layer that makes legal AI systems reliable.

‍

Task Details

• Points per question: 100
• Number of tasks: 10
• Total points available: 1000
• Project eligibility: 1,500 points

‍

Visit the app today to start earning points!

‍

Ready to Get Started?

The Multi-Issue Clause Conflict Detection Quest is now live on your Perle Labs dashboard: http://app.perle.xyz

‍

Each classification you submit helps build auditable AI infrastructure for legal systems where missing a conflict isn't an option, (not black-box pipelines), but human-verified data with a clear chain of custody.

‍

Stay Connected

Whether you're returning from beta or joining for the first time, Season 1 is open and ready. Follow along with us on X, Discord, and Telegram to stay connected so you don't miss what's coming.

‍

Human-verified. Auditable. Sovereign. Start contributing.

‍