Register to participate in the study, or sign in if you have an existing account.
Register
Sign in
Coordinator
Coordinator Dashboard
KAPS Delphi Study Management
Setup
Participants
Round 1 data
Consensus stats
Manage rounds
Settings
👥 Registered participants
Loading...
📊 Round 1 raw responses
Loading...
🎯 Consensus statistics - effectiveness matrix
Consensus: IQR of 1 or less AND 70%+ of ratings within 1 point of median. Decimal column uses band midpoints for Table 8 insertion: 1=0.05, 2=0.20, 3=0.43, 4=0.65, 5=0.88.
🎯 eD threshold consensus
💬 Qualitative responses (F1-F3)
These are open-text responses. Report as thematic narrative in the paper. No numeric aggregation should be applied.
⚙ Round management
You can open and close each round at any time. Closing a round prevents new submissions but does not delete data. You can reopen a closed round if needed.
🔄 Reset rounds
Use these before sharing the link with real participants, after testing. Resetting deletes all response data for that round and cannot be undone.
⬇ Export data
🔑 Change coordinator credentials
These credentials are stored in the Supabase settings table. After changing, you will need to use the new credentials on next login.
🔗 Supabase database connection
Update the database connection if you need to switch Supabase projects. Changes take effect after saving and reloading the page.
Round 1 Instrument
Participant
25-35 min
Section 1 of 6
👤 Section A. About you
Brief background questions to contextualise your ratings. About 3 minutes.
Sectors (select all that apply)
Jurisdictions (select all that apply)
📖 KAPS in plain language
KAPS classifies organisational knowledge assets into four sensitivity tiers and selects the minimum-cost set of safeguards that keeps residual AI risk below a tier-specific threshold and satisfies applicable regulations simultaneously.
Tier
Description
Examples
Public
No personal data; no regulatory constraint
Published reports
Personal
Standard identifiable PII (GDPR Art. 4)
CRM records, employee data
Sensitive
Special category (GDPR Art. 9) or high-value IP
HR analytics, R&D datasets
Highly Sensitive
Regulated health, biometric, financial records
NHS patient data, payment records
📏 Rating scale - used in Sections C and E
1
No / negligible effect
2
Marginal effect
3
Moderate effect
4
Substantial effect
5
Strong / near-complete
All dropdowns in Section C start blank. Please rate from your own judgment first. The small italic label below each dropdown shows the KAPS baseline for reference only - it is not a suggested answer.
📊 Section C. Safeguard effectiveness matrix
The KAPS framework defines eleven safeguards mapped against five AI-specific threat types. Your task is to rate, from your own professional experience, how effectively each safeguard reduces each type of threat under typical enterprise deployment conditions. Assume the safeguard is correctly implemented. Rate each cell independently.
⚠ Important: The small italic value shown below each dropdown (labelled "KAPS: X") is the current KAPS framework baseline value provided for reference only. It is not a suggested answer. Please form your own independent judgment before looking at the baseline. Your independent rating is the scientific value of this exercise.
All 55 dropdowns must be rated before you can proceed. Cells rated: 0 / 55
⚙ Section D. Residual-risk thresholds (eD)
The residual-risk threshold (eD) is the maximum acceptable level of risk that remains after safeguards have been applied to a knowledge asset at a given sensitivity tier. A value of 0.10 means your governance authority would only approve deployment if residual risk has been reduced to 10% or less of its unmitigated level. A value of 0.60 means 60% residual risk is considered acceptable for that tier.
⚠ Important: The KAPS baseline values shown in the third column are the current framework defaults provided for reference only. Please enter your own independent judgment in the fourth column. All four tier values must be filled in and must be between 0.00 and 1.00.
💰 Section E. Safeguard cost ranking
KAPS assigns each safeguard an ordinal implementation cost score from 1 to 5, reflecting the combined computational, operational, and economic burden of deploying that safeguard in a typical enterprise setting. The scoring is: 1 = Low (commodity tooling, standard IT operations); 2 = Low-Moderate (process overhead, moderate specialist input); 3 = Moderate (specialist expertise or infrastructure required); 4 = High (significant coordination, governance, or cryptographic overhead); 5 = Very High (specialised hardware, high compute, or complex inter-party agreements).
⚠ Important: The KAPS rank column shows the current framework defaults for reference only. For each safeguard you have deployment experience with, please indicate whether you agree with the current rank or select a revised rank. You must make a selection for every row before proceeding.
💬 Section F. Short reflections
Open-text questions. Two to three sentences per question is sufficient. About 5 minutes.
All three questions below are mandatory. If a question does not apply to your context, enter NA rather than leaving the field blank.
Round 1 submitted
Thank you
Complete
✅
Response recorded
Your Round 1 ratings have been saved. The research team will contact you at your registered email address when Round 2 opens. Round 2 returns the group distribution alongside your own ratings for review and optional revision.
Round 2 - Controlled feedback
Participant
20-30 min
Round 2 shows the group distribution from Round 1. Your own Round 1 rating is pre-filled in each row. You are not required to revise toward the group median. If your revised rating falls outside the group IQR, a brief rationale is required - this dissenting view will be reported in the paper.
Round 2 submitted
Participation complete
Complete
🏆
Thank you for participating
Your Round 2 ratings have been saved. The research team will share the paper draft incorporating the Delphi consensus values before submission.