{"resourceId":"csu-ai-systems-mou-counterproposal-aspirational-no-new-rights-20260709","versions":[{"version":"external-55c67dd34c7465da19a2d8fb584bb43a5f835c0f0c9aa59919e294bf2c652e50","resource":{"id":"csu-ai-systems-mou-counterproposal-aspirational-no-new-rights-20260709","title":"CSU's AI bargaining proposal frames faculty AI terms as non-binding principles and rejects bans on AI performance evaluation and employee monitoring","organization":"California State University (proposal to the California Faculty Association, posted by CFA)","sector":"Public university system: faculty labor relations and AI governance","geography":"California, United States","publishedAt":"9 July 2026 (document header 'CSU #3 (CFA) 07/09/2026')","publicationDate":"2026-07-09","eventDate":"2026-07-09","sourceName":"California Faculty Association (posting CSU proposal)","sourceLabel":"Collective bargaining proposal","sourceUrl":"https://www.calfac.org/wp-content/uploads/2026/07/CFA-Article-AIX-CSU-3-07.09.2026.pdf","evidenceClass":"vendor-claim","outcomeClass":"cautionary","topics":["accessibility-workforce","governance-procurement","data-security","operating-model"],"finding":"CSU's 9 July proposal places AI terms in a memorandum of understanding of 'general, aspirational principles' that does not create new rights or obligations and expires with the contract. Its text, with comments marking agreement, keeps AI from serving as Instructor of Record, leaves use or non-use of AI in instruction to faculty, requires disclosure of AI for notes and summaries, commits to vendor review and contractual safeguards for faculty data, bars AI from independently determining employment actions, bars discipline for using or refusing AI by itself, and creates a one-time $2.5 million professional development fund. Meet and confer is limited to impacts required by HEERA. CSU rejects CFA's proposed bans on AI in faculty evaluation and on AI employee monitoring, and its bar on reassigning unit work to AI.","sledRelevance":"Why now: the CFA strike authorization vote announced 22 October names AI guardrails as unresolved, and this document shows exactly where the parties differ: binding protections versus aspirational principles, and evaluation and monitoring uses.","evidence":"Employer's own bargaining proposal with margin comments; a party document that describes positions, not outcomes.","architectureImplications":"Limited; requires vendor review, risk assessment and data security standards for third-party AI services holding faculty data.","governanceImplications":"Choice between enforceable contract articles and a non-binding MOU is the central dispute; management rights to pilot or discontinue AI are reserved.","securityPrivacyImplications":"Confidential data may not be entered into generative AI in ways that compromise privacy; reasonable controls for faculty personal information and IP across cloud and LMS platforms.","caveats":"July position; parties have since moved through impasse and mediation, so current offers may differ.","streamIds":["campus-operations"],"roles":{"sales":"Sales framing: CSU's July proposal treats faculty AI terms as aspirational principles with no new rights, and rejects bans on AI evaluation and monitoring. Interpretation: institutions offering non-binding AI language may meet strong union resistance, so leaders need a clear view of which commitments they can make binding. Offer a bounded review mapping AI commitments to enforceability. Do not present this as CSU's current offer; positions after July are not public here.","engineering":"Engineering framing: the proposal agrees to vendor review, risk assessment and contractual safeguards for faculty data across cloud and learning management systems, and the disputed items concern AI evaluation and monitoring. Interpretation: pre-sales should identify any tools that score faculty performance or monitor employees, since those uses are the contested ground, and confirm vendor contracts include the data safeguards CSU already accepts.","delivery":"Delivery framing: several commitments are agreed in principle: disclosure of AI notes and summaries, no AI final employment decisions, and a $2.5 million training fund. Interpretation: delivery teams can prepare now by documenting where AI produces notes or summaries, confirming humans make employment decisions, and planning training. Avoid deploying AI in faculty evaluation or monitoring until bargaining resolves it."},"roleDetails":{"sales":{"customerProblem":"Employers proposing non-binding AI principles face unions seeking enforceable protections, especially on evaluation and monitoring.","stakeholders":["System labor relations","Provosts","CIO and information security","Faculty union leadership"],"discoveryQuestions":["Which AI commitments are binding versus aspirational?","Do any tools evaluate or monitor employees?","How will the training fund be used?","How are vendor data safeguards verified?"],"valueHypothesis":"Interpretation: clarity on enforceable commitments could narrow bargaining disputes; no outcome evidence.","potentialEngagement":"A bounded review mapping institutional AI commitments to enforceability and current tool use.","unsupportedClaims":["That this is CSU's current offer","Bargaining outcome","That non-binding terms will be accepted","Guaranteed savings"]},"engineering":{"fit":"Relevant to tools that touch faculty data, evaluation or monitoring.","architectureIntegration":"Third-party AI services and learning management systems holding faculty data need vendor review and safeguards.","prerequisites":["Vendor risk assessment","Inventory of AI features in HR and LMS tools","Data classification"],"constraints":["HEERA scope limits","Disputed evaluation and monitoring uses"],"security":"Agreed text bars entering confidential data into generative AI in ways that compromise privacy and requires reasonable controls for faculty personal information and IP.","proofOfValue":["Identify any AI scoring or monitoring of employees","Verify vendor contracts include data safeguards","Confirm AI cannot independently trigger employment actions"]},"delivery":{"work":"Prepare operations for AI commitments already agreed in principle while disputed items remain open.","dependencies":["Bargaining outcome","Tool inventory","Training program design"],"ownership":"Faculty affairs and labor relations, with IT for tool controls.","skillsAdoption":"Use of a one-time $2.5 million professional development fund if ratified.","governanceCheckpoints":["Human decision on all employment actions","Disclosure of AI notes and summaries","Hold on AI in faculty evaluation until resolved"],"acceptanceCriteria":["Proposed: inventory of AI note-taking and summary tools completed","Proposed: documented human review for any AI-assisted employment process","Proposed: training fund plan ready at ratification"],"risks":["Reliance on terms that change after impasse","Deploying contested uses","Loss of trust if principles seen as unenforceable"]}},"retrievedAt":"2026-10-23T21:48:44-05:00","enrichedAt":"2026-10-23T21:52:27-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Protections against AI replacing faculty-led instruction and against discipline for non-use; disputed limits on AI evaluation and monitoring.","procurementImplications":"Vendor review and contractual safeguards for third-party services are agreed text.","operatingModelImplications":"Bargaining over AI impacts limited to legally required scope; MOU reassessed at contract expiry.","sourceVerification":{"openedUrl":"https://www.calfac.org/wp-content/uploads/2026/07/CFA-Article-AIX-CSU-3-07.09.2026.pdf","referenceExcerpt":"This Memorandum of Understanding sets forth general, aspirational principles regarding AI systems and is not intended to create new rights or obligations.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}