Assign roles
Give independent systems separate research tasks.
Academic institutions · The model laboratory
Stop asking which model.
Start asking what the record can support.
EX MULTIS turns multi-model use into an accountable research workflow: distinct assignments, visible comparison, cited evidence, and disagreements that remain available to a researcher or instructor.
© EX MULTISInstitutional research standard
Model access becomes a method.
No black-box answer. Each research run produces a reviewable trail that can support teaching, research integrity, and responsible use policies.
Explore the academic program →Compare work
Test evidence, reasoning, agreement, and dissent.
Preserve sources
Keep claims and citations attached to their record.
Assess thinking
Evaluate a student's reasoning, not a single answer.
EX MULTIS · academic dossierEvidence recordResearch design
Source provenance
Competing readings
Methods note◉
Source provenance
Competing readings
Methods note◉
Academic institutions · The research council
One research question.
Several capable readings.
One inspectable record.
Give students and faculty a structured way to use multiple AI systems without treating any single model as the authority. The work is compared, sourced, challenged, and preserved for review.
Course research questionWill expanded transit access improve housing affordability?Scope · sources · assignment criteria
Policy analysis
+ Supply increases near transit
− Effects vary by local zoning
+ Authority: municipal data
Evidence critique
+ Published studies support impact
− Causation remains contested
+ Authority: peer-reviewed research
Source review
+ Current local record collected
− Implementation facts incomplete
+ Authority: primary documents
INTERSECTCompare
Retain dissent
Link evidence
Retain dissent
Link evidence
Qualified finding
Evidence supports improved access, with local zoning and timing retained as material conditions.
- Supporting sources
- 18
- Challenging sources
- 7
- Open questions
- 4
Academic institutions · The seminar record
Teach students how
research survives disagreement.
The research process is the learning material: source collection, independent readings, counterarguments, and a conclusion whose limits can be discussed in seminar.
Housing, transit, and public benefit
Guiding question: When does proximity to transit lower household costs?
Week 04 · Evidence workshop01Frame the question
02Gather primary and scholarly sources
03Compare independent readings
04Retain disagreements in the report
★ “A citywide effect may hide neighborhood variation.”
✦ “This claim depends on zoning being implemented.”
✧ “Evidence gap: household cost data is incomplete.”
What a student can show
- What the evidence supports
- What a capable reading challenges
- Which sources matter
- What remains uncertain