Contract Negotiation

State of play.

• The Anthropic-Pentagon contract collapse remains the defining precedent for AI vendor governance risk in federal contracting. The Department of War terminated its $200 million partnership after Anthropic refused to remove safety guardrails, designated Anthropic a "supply chain risk," and OpenAI captured the replacement contract — with Anthropic's injunction against the designation still pending (→ Anthropic and Pentagon Clash Over AI Guardrails, Leading to Contract Termination).
• In-house legal departments have seized control of AI procurement, writing measurable efficiency requirements and "AI discounts" into panel reviews and RFPs; an ACC/Everwell survey found 52% of in-house legal teams now use AI tools but only 7% have achieved measurable cost reductions — a gap that is reshaping how outsourcing deals are structured .
• Data license scope in AI contexts is headed to a federal court ruling. The Fastcase v. Alexi case — turning on whether "internal research purposes" restrictions cover AI model training and public deployment — will set the template for how courts read use-restriction clauses when a licensee's product evolves (→ Fastcase Sues Alexi Over Unauthorized AI Training Using Licensed Case Law).
• The AAA's Legal Context Protocol remains the only governance standard for AI agent transactions, addressing a structural gap as Gartner projects $15 trillion in B2B spending intermediated by AI agents by 2028 (→ AAA Launches Legal Context Protocol for AI Agent Transactions).
• For counsel advising AI vendors, enterprise technology buyers, law firms evaluating legal tech procurement, or clients with data licensing exposure, the practical baseline is: federal AI contracts carry demonstrated mid-term termination risk tied to safety governance conflicts; outsourcing agreements need active AI governance terms negotiated in — not added after signing; and "internal use" restrictions in data licenses are now litigation-tested.

Where things stand.

• AI agent transaction governance has no settled legal framework — the AAA's Legal Context Protocol is the first industry attempt to fill it. The Legal Context Protocol specifies contractual terms within a broader stack that includes payment and identity protocols; most agent-to-agent transactions currently lack verifiable terms, identified jurisdiction, or any recourse mechanism (→ AAA Launches Legal Context Protocol for AI Agent Transactions).
• Consumption-based AI pricing is replacing seat licenses across the legal tech market. Legora's Agent Pro shift follows similar moves by Anthropic and OpenAI; the trade-off — matter-level billing transparency against budget unpredictability — is now a live contract negotiation issue for firms approaching renewal .
• Federal AI contracting now carries a demonstrated supply-chain-risk designation risk. The Anthropic-Pentagon dispute — termination of a $200 million contract, supply-chain-risk designation, and a presidential order barring federal agency use — sets a template for how the administration can pressure technology vendors on AI governance without new legislation (→ Anthropic and Pentagon Clash Over AI Guardrails, Leading to Contract Termination).
• "Internal use" restrictions in data licenses are under direct judicial scrutiny. Fastcase's suit against Alexi seeks destruction of both datasets and AI model weights, treating model weights as infringing derivative works — a theory that, if accepted, would give data licensors powerful leverage to constrain AI applications of licensed material and force explicit AI training carve-outs in future agreements (→ Fastcase Sues Alexi Over Unauthorized AI Training Using Licensed Case Law).
• AI outsourcing governance is a recognized gap in existing vendor agreements. An ACC/Everwell survey found 52% of in-house legal teams use AI tools but only 7% have achieved measurable cost reductions; the ACC updated its AI Center of Excellence toolkit for 2026, but performance metrics tied to AI implementation and regulatory obligations remain undefined across most current outsourcing arrangements .
• Proprietary AI contract review is emerging as a competitive differentiator. Shoosmiths' Project Apollo, built on Microsoft Azure over twelve months, embeds firm-specific playbooks and audit trails into contract review — positioning institutional knowledge as a protected asset rather than licensing it to third-party vendors .
• MLS feed access disputes are testing the boundary between contract enforcement and antitrust. The Zillow-MRED-Compass dispute — a federal antitrust suit, temporary injunction restoring feeds, and unresolved claims over whether MRED's cutoff was legitimate contract enforcement or anticompetitive coordination — will shape how data licensing agreements in platform markets are drafted and enforced .
• Public sector labor contract negotiations are at an inflection point in New York City. Brooklyn Defender Services and four other ALAA locals are operating under expired contracts, with a strike authorization vote clearing at 96% at BDS — a pattern that signals systemic compensation structure disputes across public defense organizations .

Latest developments.

• Fastcase v. Alexi heads to federal court in D.C. on whether "internal research purposes" restrictions in a 2021 data license cover AI model training and public deployment; Fastcase seeks destruction of model weights as infringing derivative works (→ Fastcase Sues Alexi Over Unauthorized AI Training Using Licensed Case Law)
• ACC/Everwell survey data shows 52% of in-house legal teams use AI tools but only 7% have achieved measurable cost reductions; Irina Beschieriu identifies the structural fix as governance frameworks with AI-specific performance metrics negotiated into outsourcing agreements at signing
• Brooklyn Defender Services union votes 96% to authorize a work stoppage; four other ALAA locals operating under expired contracts; BDS is the second-largest public defense organization in New York City

Active questions and open splits.

• "Internal use" restrictions and AI model training — what the contract actually covers. Fastcase v. Alexi will determine whether a licensee's pivot from human-reviewed services to AI-driven platforms constitutes a material breach of use restrictions, and whether model weights trained on licensed data are derivative works subject to destruction — the answer reshapes drafting standards for every data license that predates the generative AI era (→ Fastcase Sues Alexi Over Unauthorized AI Training Using Licensed Case Law).
• Supply-chain-risk designation as a contract termination mechanism. The Anthropic precedent leaves open whether a supply-chain-risk designation survives judicial review, what procedural protections attach before designation, and how AI vendors should draft force majeure, termination-for-convenience, and governance-change clauses in federal contracts to account for this risk (→ Anthropic and Pentagon Clash Over AI Guardrails, Leading to Contract Termination).
• AI governance terms in outsourcing agreements — what the contract must specify. The ACC survey gap — widespread AI adoption, negligible cost reduction — points to agreements that lack performance metrics, audit rights, and regulatory compliance obligations tied to AI implementation; no market-standard terms have emerged for how these provisions should be structured .
• Enforceability of AI agent transactions under existing contract law. The Legal Context Protocol is an industry attempt to fill a gap that courts and regulators have not yet addressed — whether LCP-embedded terms will be treated as enforceable agreements, and under which jurisdiction's law, is entirely open (→ AAA Launches Legal Context Protocol for AI Agent Transactions).
• Consumption-based AI pricing — what goes in the contract. Legora's per-run model and the broader shift away from seat licenses leave measurement methodologies, audit rights, cost caps, and dispute resolution over billable-unit calculations undefined in most current agreements; no market-standard terms have emerged .
• MLS data licensing — contract enforcement versus antitrust. The Zillow-MRED dispute turns on whether a feed cutoff for non-compliance with display policies is legitimate contract enforcement or anticompetitive coordination; the outcome will define how exclusivity and compliance provisions in data licensing agreements can be enforced in platform markets .
• AI safety governance clauses in commercial contracts. The Anthropic-Pentagon dispute crystallizes a question that will recur across defense and civilian government contracts: whether a vendor's non-negotiable safety conditions can be drafted as enforceable contract terms rather than aspirational policy, and what remedies attach if the counterparty demands their removal (→ Anthropic and Pentagon Clash Over AI Guardrails, Leading to Contract Termination).

What to watch.

• The D.C. federal court ruling in Fastcase v. Alexi — if the court accepts the model-weights-as-derivative-works theory, every data license with pre-AI "internal use" language becomes a renegotiation or litigation candidate (→ Fastcase Sues Alexi Over Unauthorized AI Training Using Licensed Case Law).
• Whether Anthropic's injunction against the supply-chain-risk designation survives — the ruling will define the procedural and substantive limits of this enforcement mechanism for all AI vendors with federal contracts (→ Anthropic and Pentagon Clash Over AI Guardrails, Leading to Contract Termination).
• Whether the Brooklyn Defender Services strike materializes and whether it triggers contract settlements across the four other ALAA locals operating under expired agreements — the compensation benchmarks set here will ripple through public sector legal services contracting in New York .
• Whether Legora's consumption-based pricing model prompts market-standard contract terms for per-run billing — audit rights, cost caps, measurement definitions — to emerge from the renewal conversations now underway at major firms .
• Substantive proceedings in the Zillow federal antitrust case — whether MRED's feed cutoff is characterized as contract enforcement or anticompetitive coordination will reshape data licensing drafting in platform markets .
• Whether additional AI vendors in federal contracts face supply-chain-risk designation pressure and whether the Anthropic model — resuming talks without retreating from original safety conditions — becomes the template for vendor response (→ Anthropic and Pentagon Clash Over AI Guardrails, Leading to Contract Termination).