Our Mission

To foster collaboration between students, researchers, associations and enterprises to develop AI solutions with positive social and environmental impact. We support, develop and disseminate AI projects that respect ethics, human dignity and the environment.

Why Now?

We are at a pivotal moment. The global AI market is projected to reach $1.3 trillion by 2032 at over 37% annual growth. Yet AI is advancing faster than our ethical frameworks and social capacity to manage it. Toulouse Hub exists to close that gap — locally first, globally in reach.

Our Three Pillars in Practice

• Responsibility — every project undergoes an Ethical Impact Assessment applying the EU AI Act risk framework and UNESCO AI Ethics Recommendation.
• Innovation — structured ideation programmes, hackathons and challenges channelling creativity toward measurable social and environmental outcomes.
• Cooperation — our network spans INSA Toulouse, Université Toulouse III, CHU, local NGOs, regional enterprises and international AI for Good networks.

Who We Serve

• Young talent, students and early-career researchers in AI and related fields
• NGOs and associations seeking AI tools to amplify their social mission
• Schools, lycées and student clubs exploring responsible technology
• Entrepreneurs and innovators across the Occitanie region
• Companies committed to ethical AI and ESG-aligned technology

Aligned with: UN AI for Good framework; UNESCO Recommendation on the Ethics of AI (2021); EU AI Act (2024); ITU AI for Good.

Grounding Frameworks

• UNESCO Recommendation on the Ethics of AI (2021) — the first global normative instrument, adopted by 193 member states. We apply its four core values: respect for human rights, environmental flourishing, diversity & inclusiveness, and peaceful societies.
• EU AI Act (2024) — the world's first comprehensive AI regulation, in phased application from February 2025. We classify all supported projects by risk level ensuring high-risk applications meet full transparency and human oversight requirements.
• ITU AI for Good framework — guiding our focus on AI applications contributing to the UN Sustainable Development Goals.
• OECD AI Principles — providing our baseline for inclusive growth, sustainable development and well-being.

Our Ethical Review Process

• Bias & Fairness Audit — identifying discriminatory outcomes in training data, model architecture and deployment context
• Transparency Review — can affected users understand why a decision was made?
• Privacy & Data Governance — GDPR compliance, data minimisation, consent mechanisms and retention policies
• Environmental Impact Assessment — is the computational cost proportionate to the social benefit?
• Inclusion Check — does the solution narrow existing inequalities rather than widen them?

Sources: UNESCO AI Ethics (2021); EU AI Act (Regulation 2024/1689); ITU AI for Good; OECD AI Principles; Bletchley Declaration.

What We Provide

• Needs Analysis & Goal Definition — defining a realistic AI objective aligned with your mission — the most useful solution, not the most technically impressive one.
• Ethics & Social Impact Advising — structured ethical review applying the EU AI Act risk framework and UNESCO AI Ethics principles at every stage.
• AI Literacy Workshops — bespoke learning sessions on AI foundations, responsible data practices and tools relevant to your domain.
• Proof of Concept Development — rapid prototyping using the most appropriate current tools — from open-source foundation models to lightweight edge AI.
• Technical Orientation — guidance on AI frameworks, datasets, deployment platforms and monitoring systems for long-term project sustainability.

2025 AI Tools We Work With

• Large Language Models — Mistral, LLaMA 3, Phi-3 for NGOs and associations with limited compute budgets
• Computer Vision — YOLO v10, SAM 2 (Segment Anything Model) for environmental monitoring
• Responsible AI Toolkits — IBM AI Fairness 360, Google What-If Tool, Microsoft Responsible AI Toolbox
• Edge Deployment — TensorFlow Lite, ONNX Runtime for IoT and mobile-first applications

Workshop Catalogue

• Introduction to AI — what machine learning actually is, how models are trained, what they can and cannot do. No prior technical knowledge required.
• AI Ethics & Security — algorithmic bias, transparency, fairness, data protection (GDPR), and the social consequences of high-stakes AI deployment.
• AI for Climate & Environment — biodiversity monitoring, energy grid optimisation, climate modelling and sustainable agriculture applications.
• Practical AI Labs — build a working chatbot, train a simple image classifier, or create a mini-application using no-code and low-code AI tools.
• AI Awareness in Education — tailored programmes for secondary schools and youth organisations: age-appropriate introduction to AI's capabilities, risks and societal implications.

Why AI Literacy Is Urgent in 2025

By 2030, an estimated 70% of jobs will require significant AI interaction. The EU's Digital Compass targets 80% of adults with basic digital skills by 2030 — currently only 54% have them. The proliferation of generative AI (ChatGPT, Gemini, Copilot) makes foundational AI literacy a civic necessity, not just a career skill.

How Open Innovation Works at the Hub

• Project-Partner Matching — connecting project holders with AI expertise, domain knowledge and community relationships they need to succeed.
• Hackathons & Challenges — intensive co-creation events structured around defined social or environmental challenges: urban waste, digital inclusion, refugee integration support.
• Rapid Prototyping Support — technical and methodological scaffolding to move from challenge definition to working prototype in weeks, not months.
• Mentorship & Follow-Through — structured mentorship tracks transforming hackathon prototypes into deployable solutions.
• Open Publication — methodologies, datasets and results shared openly, contributing to the global AI for Good knowledge commons.

2025 Open Innovation Trends We Apply

• Federated Learning for NGOs — training shared AI models without centralising sensitive beneficiary data
• Foundation Model Fine-Tuning — using Mistral and LLaMA 3 as bases for domain-specific tools, reducing cost and time dramatically
• Human-AI Collaborative Design — participatory design ensuring tools are built with communities, not just deployed at them
• Impact Measurement Frameworks — Theory of Change models ensuring measurable outcomes are defined before development begins

About the AI for Good Global Summit

The AI for Good Global Summit is the ITU's (International Telecommunication Union) flagship annual gathering held in Geneva, Switzerland. It brings together AI researchers, UN agency leaders (WHO, UNICEF, UNEP, UNESCO), government policymakers, industry innovators and civil society — united by one mission: to identify practical AI applications to accelerate the 17 UN SDGs.

Summit 2026 Programme Areas

• AI for Good Impact Awards — recognising the most impactful AI projects advancing the SDGs globally
• Canvas of the Future — visionary sessions on long-term AI trajectories and societal implications
• AI for Good Film Festival — storytelling at the intersection of AI and human experience
• AI Skills & Capacity Building — global programmes for AI literacy, access and education equity
• AI Standards & Policy — shaping international regulatory frameworks governing AI deployment
• Young AI Leaders Programme — ITU's structured pathway for the next generation of responsible AI leaders

Adam's Volunteer Journey (Official Profile)

• Noah's Ark Technology Digital Solidarity (from May 2024) — technological assistance, workshops and tailored digital support to reduce the digital divide, "demonstrating the transformative power of technology in fostering equality and social progress."
• SNESC — Solidarité Numérique, Éducation et Santé Castres (from June 2023) — community event organisation, digital literacy campaigns and local initiatives for underserved populations.
• Noah Production Genève — simulated CEO role in a volunteer initiative — developing leadership, project management and strategic planning skills for global challenge-solving.

Source: Official ITU AI for Good Global Summit speaker profile — Adam Kouadri.

What is Davos House AI?

Davos House AI brings together experts, policymakers, researchers, entrepreneurs and civil society actors from around the world, held alongside the World Economic Forum in Davos, Switzerland. Its objective: to collectively shape how AI can address global challenges — climate change, education, health, digital inclusion, governance and peace.

The Toulouse Hub's Contribution

• Human-centred AI — Advocating for AI development frameworks that place dignity, autonomy and cultural diversity at the centre of design and deployment decisions.
• Reducing the digital divide — Presenting inclusive technology models from Toulouse — demonstrating how regional hubs can serve underserved populations with minimal compute resources and maximum social impact.
• AI for the ecological transition — Showcasing Smart Waste & Recycling AI as a proof of concept for city-scale AI deployments generating measurable CO₂ reductions and circular economy benefits.
• Local hubs as global catalysts — Making the case that AI for Good does not require Silicon Valley-scale resources. Regional hubs, rooted in local realities, can produce globally replicable solutions.

What Opens Next

• Development of international AI for Good partnerships with organisations represented at Davos
• Launch of cross-border AI for Good pilot projects building on the Smart Waste methodology
• Strengthening Toulouse Hub's role as a European reference point within the ITU/UNESCO AI for Good network

Source: Toulouse Hub AI for Good — Official Davos House AI 2026 Participation Report.

The Challenge

Toulouse spends €25 million per year on waste management, achieves only a 65% recycling rate and generates 500 tonnes of avoidable CO₂ annually from inefficient collection routing. Standard waste systems operate on fixed schedules with no intelligence — collecting empty bins and missing full ones.

Our Solution Architecture

• 10,000 IoT Smart Bin Sensors — ultrasonic fill-level monitoring + NIR spectrometers for real-time composition analysis, connected via LoRaWAN
• LSTM Time-Series Forecasting — predicts waste volume peaks 24–72 hours ahead using historical fill-rates, weather, events and seasonal patterns
• Q-Learning Route Optimisation — reinforcement learning agents dynamically rerouting the collection fleet, cutting km driven by 20%
• Blockchain Citizen Incentives — Hyperledger Fabric recording verified sorting actions, converting them to tax reductions (up to €85/household/year)
• Real-Time Dashboard — Kafka → Spark Streaming → Power BI with contamination alerts and CO₂ accounting

UN SDG Alignment

• SDG 11 — Sustainable Cities: smart urban infrastructure reducing waste and improving quality of life
• SDG 12 — Responsible Consumption: circular economy enablement through AI-powered material recovery
• SDG 13 — Climate Action: measurable IPCC-verified CO₂ reduction with carbon credit generation

Sensing & Connectivity

• Ultrasonic + LiDAR fill-level sensors; NIR spectrometers for material identification; RFID household tags
• LoRaWAN / NB-IoT for fixed bin network; 5G for vehicle telemetry and sorting centre video streams

Edge AI — MobileNetV3

INT8-quantized CNN models embedded in bin cameras for local waste pre-classification. MDPI Sensors research (June 2025) validates this architecture for IoT-scale waste classification at the point of disposal — enabling offline inference with minimal power consumption.

Cloud AI — ResNet-152 + Vision Transformer

• 500k+ image training corpus, 50+ material categories. Hybrid CNN-ViT ensemble achieves best-in-class accuracy on benchmark waste datasets as of 2025
• LSTM networks for 72-hour waste volume forecasting
• Q-Learning agents for multi-objective route optimisation (fill urgency, fuel cost, CO₂ budget)
• GAN data augmentation for synthetic rare-class training data

Data Infrastructure

Apache Kafka → Spark Streaming → Feast Feature Store → TensorFlow Serving. Power BI + ELK Stack for real-time operator dashboards and public-facing reporting.

Blockchain & Privacy

• Hyperledger Fabric for material traceability and citizen incentive management
• Homomorphic encryption for sensitive waste data processing
• Federated learning for privacy-preserving model training across facilities

Research: MDPI Sensors (June 2025); Columbia Climate School (2025); Resource Recycling (Dec 2025); AMP Robotics; Greyparrot AI.

Digital Literacy Programmes

We design and deliver workshops, training sessions and awareness campaigns tailored to the digital skills gap in local communities — in collaboration with associations, schools and local stakeholders across the Occitanie region. Inspired directly by Adam's volunteer work with Noah's Ark Technology Digital Solidarity and SNESC.

Equitable Access to AI

We specifically target underserved and marginalised populations — seniors, youth in under-resourced neighbourhoods, first-generation university students — ensuring the opportunities created by AI are accessible to all, not just those already advantaged by education and economic privilege.

2025 AI Literacy Context

• The EU's Digital Compass targets 80% of EU adults with basic digital skills by 2030 — currently only 54% have them
• The proliferation of generative AI tools (ChatGPT, Gemini, Copilot) makes foundational AI literacy a civic necessity
• UNESCO's 2023 guidance for schools on generative AI emphasises critical thinking about AI-generated content, not just technical proficiency

Ethical Awareness in Education

Our programmes are aligned with UN and ITU frameworks, with strong emphasis on transparency, fairness and data protection. Students learn not just how to use AI tools, but how to question them, challenge their outputs and understand their societal implications.

Our Approach

The Hub's health work focuses on strengthening social infrastructure and inclusion through AI — rather than attempting to develop complex medical technologies (drug discovery, diagnostic imaging) which are appropriately led by large, regulated research institutions.

Technological Assistance

We provide tailored digital support to individuals and local organisations navigating health and social services — including support for digital health tools, health information literacy and AI-assisted navigation of public service systems.

Local Partnerships

• CHU de Toulouse — digital health literacy and ethical AI use in hospital settings, focusing on patient information rights and data protection
• Toulouse School of Economics — research on health economics applications, including AI for optimising public health resource allocation in Occitanie

2025 AI in Health Context

The WHO issued guidance in 2024 on the ethics and governance of large multimodal AI models in health. Key principles we apply: safety, effectiveness, equity, accountability, sustainability and transparency — ensuring AI health tools benefit all patients, not just those in well-resourced settings.

Current Focus: Smart Waste & Recycling AI

Our flagship environmental project applies IoT, LSTM forecasting and Q-Learning to Toulouse's waste system — targeting −40% landfill waste, +30% recycling rates and −15% CO₂ emissions in Year 1.

Future Environmental AI Initiatives

• Urban Green Space Monitoring — computer vision + satellite imagery for biodiversity and urban heat island tracking
• Energy Grid Optimisation — AI demand forecasting for Toulouse's municipal buildings through intelligent HVAC and lighting control
• Circular Economy Facilitation — AI matching platforms connecting industrial waste producers with potential reusers across the Occitanie ecosystem
• Climate Risk Modelling — AI models for urban flood risk and heat stress prediction under various climate scenarios (with INSA Toulouse)

AI & Climate — 2025 State of Play

The IPCC Sixth Assessment Report identified AI as a cross-cutting enabling technology for climate mitigation and adaptation. Key precedents: DeepMind's AlphaFold accelerating biofuel research; Google's fuel-saving AI for air traffic control; ML models reducing the cost of solar and wind power forecasting by up to 35%.

Our Advocacy Work

• International forums — Davos House AI 2026, ITU AI for Good summits and European AI governance consultations, ensuring civil society and regional perspectives are represented
• Local policy engagement — contributing to Toulouse Métropole's digital strategy consultations with AI expertise for smart city technology deployment
• Academic collaboration — joint research with INSA Toulouse and Toulouse School of Economics on AI governance frameworks, published openly

Key 2025 Governance Developments We Track

• EU AI Act — phased application from February 2025. We translate risk classifications into practical guidance for Hub partners
• G7 AI Code of Conduct — voluntary commitments on responsible AI development used as reference benchmark for evaluating commercial AI partnerships
• UNESCO AI Ethics Recommendation implementation — tracking which of 193 signatory states are converting the Recommendation into binding national policy
• Global AI Safety Summit — Bletchley Declaration — principles incorporated into our risk assessment frameworks for high-capability AI projects

ITU Young AI Leaders — Official Profile

• Role: Hub Leader
• Period: 20 January 2025 – 2029
• Hub Name: Toulouse Hub
• Region: Europe  ·  Country: France
• Organisation: Noah's Ark Technology Digital Solidarity
• Profession: Project Coordinator

Volunteer Journey

• Noah's Ark Technology Digital Solidarity (from May 2024) — technological assistance, workshops and tailored digital support to reduce the digital divide, demonstrating "the transformative power of technology in fostering equality and social progress."
• SNESC — Solidarité Numérique, Éducation et Santé Castres (from June 2023) — community events, digital literacy campaigns and local initiatives for underserved populations.
• Noah Production Genève — simulated CEO role in a volunteer initiative, developing leadership and strategic planning skills for global challenge-solving.

Smart Waste & Recycling AI

Adam is also Founder & President of Smart Waste & Recycling AI — the Hub's flagship applied project targeting −40% landfilled waste, +30% recycling rates and −15% CO₂ in Toulouse through IoT, LSTM forecasting and blockchain-based citizen incentives.

Core Philosophy

"AI should not be a tool of exclusion or a source of new inequalities. It should be, above all, a force for justice, sustainability and human dignity."

Partnership Models

• Academic Partnerships — Joint research, dataset sharing, student placement and co-publication. Current partners: INSA Toulouse, CHU de Toulouse, Toulouse School of Economics.
• Corporate Partners — Responsible deployment frameworks, AI literacy training and SDG-aligned technology projects for companies integrating ethical AI practices.
• NGO & Association Networks — Connecting social impact organisations with the AI capabilities they need to amplify their missions, without requiring in-house technical expertise.
• International AI for Good Networks — Cross-border partnerships with AI for Good hubs in Europe, Africa and Asia, following our Davos 2026 participation.
• Public Sector — Guidance, co-development and independent ethical review for municipalities and public institutions deploying AI in alignment with the EU AI Act.

What We Offer Partners

• Access to our community of AI researchers, students and practitioners in Toulouse
• Ethical AI review and responsible deployment frameworks
• Co-branding and visibility within the international AI for Good ecosystem
• Joint grant applications for EU Horizon Europe and AI for Good funding streams
• Participation in our open innovation events and hackathons

📧 toulousehub.youngaileaders@gmail.com

👤 Adam Kouadri — Hub Leader & Founder

📍 Toulouse, Occitanie, France

Who Should Join?

• Students & young researchers — at any level, in any field. AI for Good is interdisciplinary: we need economists, sociologists, designers and artists as much as engineers.
• AI practitioners — data scientists, ML engineers, software developers who want to apply their skills to problems that matter.
• Social innovators — NGO workers, social entrepreneurs and community organisers who need AI expertise to amplify their missions.
• Educators & trainers — teachers and facilitators interested in delivering AI literacy programmes or collaborating on educational content.
• Professionals & companies — organisations committed to responsible AI practices looking for a community and validation framework.

What Membership Gives You

• Access to all Hub workshops, training sessions and learning resources
• Participation in hackathons, challenges and open innovation events
• Connection to our network of researchers, practitioners and partner institutions
• Mentorship opportunities — as a mentee or mentor
• The opportunity to contribute to and lead Hub projects with real social impact
• Visibility within the international AI for Good ecosystem, including through ITU Young AI Leaders and Davos partnerships

📧 toulousehub.youngaileaders@gmail.com

👤 Adam Kouadri — Hub Leader

Include your name, background, area of interest and what you hope to contribute to or gain from the Hub community.