Our AI ethics framework is not a compliance document. It is a set of architectural commitments that shape how every system on the platform is built, deployed, and constrained — for every human, every organization, every mission.
The most dangerous AI systems are the ones that make consequential decisions without telling anyone. GameChangers was built with the opposite principle: every AI action is logged, every AI decision is explainable, every AI permission is explicitly granted by a human, and every AI authority has a hard boundary.
This is not a limitation. It is the design. The professionals who use this platform are working on defense systems, health infrastructure, space programs, and energy networks. The stakes are too high for opaque AI behavior.
Every Curious PI recommendation includes the reasoning that produced it. Every Trust Score shows exactly which sources contributed what. A professional can always ask "why" and receive a specific, verifiable answer.
Every new scoring formula is reviewed for proxy discrimination before it reaches production. "Years of experience" as a proxy for competence, geographic location as a signal for network quality — these are examples of signals we actively evaluate and reject when they encode demographic bias.
Demographic attributes including underrepresented status are opt-in only. They are encrypted at the application layer using AES-256 before storage. They are never used in matching algorithms without explicit user consent and documented legal review. They can be deleted at any time from Settings.
GameChangers has no access to DCSA adjudication databases. No commercial platform does. When a professional indicates a clearance level, it is displayed as "Self-certified — not platform-verified" in every context. The label is architectural, not optional. Misrepresenting clearance status violates our Terms of Service.
The platform flags ITAR compliance requirements on defense and space missions. It surfaces the flag as information for the mission lead. The platform never makes ITAR compliance determinations — those are legal decisions that require a licensed attorney. AI-assisted contract vetting identifies potential gaps; humans and legal counsel determine compliance status.
When a professional enables Mission Intelligence Agents on a contract, the following clause appears in the contract automatically and cannot be removed:
Agents can only do what the Mission Lead explicitly enables. Default state: all five gates are off. The gates are: (1) auto-respond to inquiries, (2) vet incoming contracts, (3) post mission updates, (4) invite members, (5) manage calendar availability. No gate exists for financial approval, contract signing, or dispute resolution — those remain human-only.
Every permission gate has a Redis rate limiter. Auto-respond: 20/hour. Contract vetting: 10/hour. The rate limits are not just cost controls — they enforce a pace of action that keeps agents within the review capacity of the human overseeing them. An agent that acts 500 times per hour cannot be meaningfully supervised. 20 times per hour can.
The verification system was designed to be geographically equitable. A researcher at IIT Bombay with 30 ORCID publications and an ISRO grant has a credible path to a full Trust Score — not through US-centric signals, but through the sources relevant to their work and region.
US: ORCID + SAM.gov + Semantic Scholar + USPTO patents
EU: ORCID + Scopus + Dimensions + VIES (VAT)
Brazil: ORCID + Lattes Platform (CNPq) + Semantic Scholar
India: ORCID + Semantic Scholar + Scopus + GSTIN (org)
Global: ORCID + Semantic Scholar + arXiv + IEEE + Dimensions
We do not use: the prestige of the professional's institution, their geographic location, their name (no language/origin inference), their educational institution's global ranking, or any inferred demographic signal as a factor in matching ranking or Trust Score.
The engineer in Lagos deserves the same shot as the one at Lockheed. This is not a marketing line — it is an architectural constraint. Every system that would disadvantage a qualified professional based on where they live, who they know, or what institution they attended is a system we review and reject.
We're building in public. If you have questions about our AI architecture, ethics decisions, or verification methodology, reach out directly.
Contact AI Team →