Open cohorts
Learn through Eleni’s world
A shared fictional workplace gives every learner realistic evidence, decisions and constraints without exposing real organisational data.
See dates and cohort detailsResponsible AI in practice
Responsible AI training for complex, accountable work.
Build practical AI capability through live practice, visible safeguards and professional judgement. Join an open cohort in Eleni’s fictional workplace, or commission a programme adapted to your organisation.
Live facilitation. Synthetic practice. Accountable human decisions.
Choose the route that fits
Join a shared open cohort or commission a programme adapted to your organisation.
Custom programmes
Built around your audience
Adapt the roles, source material, tasks and governance context for an agency, government team or organisation while keeping the AidGPT core.
Discuss a tailored programme
AI Disclosure and Use (MI-GUIDE-001)
A free ten-page guide for managers: what staff may use AI for, what must be disclosed, and a staff notice ready to adapt and issue. Free, ungated and CC BY 4.0.
AidGPT Updates is our email list on practical AI adoption, training and governance.
Sign upTrusted by
Choose your format
Responsible AI in Practice
Practical training for aid, humanitarian, development, social-protection, governance, donor, and public-sector professionals.
Practise
Live Intro Lab
A facilitated introduction to responsible AI workflows.
- Format
- 90-minute facilitated group session
- Investment
- EUR 75
- Workflow Card
- Lite Workflow Card
- CHEF method
- Verification prompt practice
- One approved low-risk task mapped into a card
Build
Full six-session open cohort
Move from one-off prompting to responsible, repeatable workflows.
- Format
- Six 90-minute live sessions over three weeks
- Investment
- EUR 350 per person EUR 280 reduced rate for eligible participants — national and local NGO staff in low- or middle-income or crisis-affected countries, and aid workers between roles
- Workflow Card
- Full Cohort Workflow Card
- Reusable Project or agent design
- Verification prompts
- Certificate of completion
Embed
Private team cohort
The full cohort adapted to your team workflows and risk boundaries.
- Format
- Six-session structure adapted to your organisation
- Investment
- Custom quote for teams
- Workflow Card
- Full Cohort Workflow Card
- Team-specific examples
- Shared workflow cards
- Optional post-course advisory
What learners and partners say
Hear from humanitarian professionals who have used AidGPT to move from curiosity to practical, responsible AI work.
"I signed up a little nervous it might be too introductory. It wasn't. The instructors pitched it at a useful technical level for people already working with AI in practice. The cohort format meant real-time pressure-testing of ideas, and other participants' questions sharpened my thinking. If you work in the humanitarian or development sector and want to go beyond the hype on AI, this is worth your time."
Dan Amias
Elrha
"The training took something complex and made it workable. The trainers have a rare gift for translating daunting territory into practical, usable guidance that lands across different roles and industries. I walked in uncertain. I walked out confident, equipped to use these tools wisely, and clear-eyed about what I should and shouldn't do."
Leah Donoghue
AidGPT June 2026 cohort
"This training helped our teams understand AI in a practical, responsible way. People came out with real workflows they can use immediately, clearer safeguards, and the confidence to work safely with these tools. It was one of the most relevant and useful trainings we've had this year."
Gaurav Singh Basnyat
Digital Programme Development Manager, Norwegian Refugee Council
Practical capability
Practise the work that responsible AI requires
By the end of the programme, learners should be better able to:
Brief and challenge
- brief one meaningful task with enough context and a clear standard of quality
- challenge a first answer and show what was checked
Save and connect
- save useful context and reusable working patterns deliberately
- use connected capabilities with bounded reach and explicit stop decisions
Work and decide
- work across files with checkpoints and ownership
- decide what should be adopted, governed or left alone
A controlled synthetic workplace for advanced AI practice.
AidGPT is developing a visible training environment where learners can practise more capable AI workflows using fictional records. Publication of connected features follows product acceptance testing.
- Product state
- Pre-acceptance
- Data boundary
- The course does not require real organisational or personal data.
Learn through Eleni’s working world. Apply the discipline to yours.
Eleni Andreou is Head of Programmes at Koinon Relief Partnership in the fictional Republic of Alashiya. She works across programme priorities, evidence, partners and operational constraints during early recovery. Her role gives the cohort a consistent point of view without using real organisational or personal information.
Synthetic course records
- Fictional role
- Eleni Andreou Head of Programmes
- Fictional organisation
- Koinon Relief Partnership
- Fictional setting
- Republic of Alashiya
For organisations
One core learning structure, grounded in your people’s work.
Commission an AidGPT programme adapted to your roles, information environment, tools and risk boundaries. Keep the responsible AI disciplines while changing the context that makes them relevant.
- What stays fixed
- Responsible AI disciplines, verification, human judgement, synthetic practice and reflection.
- What adapts
- Roles, setting, documents, tasks, tools, policy and risk boundaries.
Examples include government teams, development agencies, aid and development organisations and businesses.
Discuss a tailored programmePermission-cleared evidence
What learners and partners say
A representative selection from AidGPT training.
I signed up a little nervous it might be too introductory. It wasn't. The instructors pitched it at a useful technical level for people already working with AI in practice. The cohort format meant real-time pressure-testing of ideas, and other participants' questions sharpened my thinking. If you work in the humanitarian or development sector and want to go beyond the hype on AI, this is worth your time.
The training took something complex and made it workable. The trainers have a rare gift for translating daunting territory into practical, usable guidance that lands across different roles and industries. I walked in uncertain. I walked out confident, equipped to use these tools wisely, and clear-eyed about what I should and shouldn't do.
This training helped our teams understand AI in a practical, responsible way. People came out with real workflows they can use immediately, clearer safeguards, and the confidence to work safely with these tools. It was one of the most relevant and useful trainings we've had this year.
Common questions
What to know before you choose
Do I need to bring my own data?
- No. Assessed Glass Box practice uses synthetic records. Real organisational or personal data is not required.
Is the open cohort only for humanitarian work?
- The open cohort uses Eleni’s aid and development setting because it creates a coherent, consequential practice world. AidGPT also offers customised programmes for government, development agencies, organisations and businesses.
Can you adapt the course to our organisation?
- Yes. A custom programme can adapt roles, source material, tasks, tools and governance context while retaining the responsible AI disciplines.
Does the Glass Box connect to our production systems?
- No production connection is required. The Gen 5 practice model uses synthetic training services. Any later use of organisational data follows the organisation’s own approved route.
Is MCP what makes it secure?
- No. MCP is a connection protocol. Security and safety depend on the accepted product controls, permissions, isolation, operating rules and human judgement.
Will the website show the full curriculum?
- The site explains the experience, outcomes and safety boundary. The source pack, full task design and facilitation method remain inside the programme.