Build the community
We create a safe space where people working on AI can meet, ask difficult questions, share experiences, and learn from each other.
Non-profit · community-run · open source
Responsible AI Movement
Responsible AI Movement is a non-profit global collective of researchers, designers, and AI practitioners on a mission to make AI ethics more practical.
We are collectively defining what Responsible AI means and sharing practical tools with AI designers and developers to create responsible Conversational AI products: AI chatbots, voice agents, and digital humans.
What we do
We bring different perspectives together.
We bring together designers, developers, researchers, product people, ethicists, and others working with AI to look at the same problems from different angles.
Together, we research risks, discuss real-world cases, challenge assumptions, and work towards a fuller picture of what responsible AI can look like in practice.
Do the research
We dig into scientific research and real-life cases to understand what can go wrong and how those risks show up in AI products.
Make it practical
We turn what we learn together into open resources, workshops, questions, and mitigation strategies that teams can use.
Spread the word
We create spaces for people to discuss responsible AI through workshops, talks, events, and community conversations.
The Responsible AI Launch Kit
Practical tools for responsible AI teams.
Responsible AI can feel abstract. We want to make it something teams can actually work with.
For each risk zone, we dig into real-life cases of what's the worst that can happen, read scientific research, define questions to help teams spot the risk, and put together must-have mitigation strategies.
The eight AI Ethics risk zones
Z.01
Data privacy
Potential for an AI system to collect, process, store, or expose personal or sensitive information in ways that harm individuals or organisations, whether through misuse, inadequate protection, unintended disclosure, or lack of meaningful consent and control.
Z.02
Disinformation
Potential for an AI system to generate, amplify, or present false, misleading, or unverified information as credible and authoritative, whether through hallucination, synthetic media, or deliberate manipulation, in ways that mislead individuals, distort decision-making, and cause harm to people, organisations, and society at large.
Z.03
Addiction and emotional dependency
Potential for an AI system to encourage excessive user reliance, through emotional dependency, persuasive design, cognitive offloading, or decision delegation, in ways that erode autonomy, weaken user capabilities, and negatively affect long-term wellbeing.
Z.04
Algorithmic bias
Possibility that an AI system makes systematically unfair decisions that harm individuals or violate their rights, through unrepresentative training data, flawed model design, or unnoticed reinforcement of existing inequalities, narrowing access to housing, finance, education, insurance and employment for the people who need them most.
Z.05
Bad actors
Possibility that chatbot platforms are deliberately misused for harm, whether by exploiting general-purpose assistants to mass-produce convincing scam messages, or by using chatbot development platforms for building purpose-made bots that defraud, launder money, impersonate trusted institutions or manipulate users at a scale no human operation could sustain.
Z.06
Environmental risk
Potential for an AI system to consume energy, water, and natural resources, or generate emissions and waste, in ways that harm ecosystems, communities, or the climate, whether through inefficient design, unsustainable hardware and infrastructure choices, or lack of transparency and accountability for environmental impact.
Z.07
Loss of human control
Potential scenarios in which AI systems pursue goals or take actions that humans can no longer meaningfully oversee, correct or override, whether because the AI has become too capable, too autonomous or too deeply embedded in critical systems for humans to intervene effectively.
Z.08
Loss of independent thinking
Erosion of people's capacity to reason, judge and decide for themselves, through over-reliance that atrophies underused skills, uncritical acceptance of fluent output, and the narrowing of viewpoints when a handful of models mediate what most people read, write and believe.
The resource is completely free. It includes:
- 8 AI ethics risk zones with questions to help your team spot which tech risks your product has and define mitigation strategies.
- The workshop playbook to help your team run a responsible AI workshop.
- An example responsible AI workshop your team can run.
Join the community
A space to ask the uncomfortable questions.
If you work in AI and want to meet people from different backgrounds, share what you're seeing, and think through difficult questions together, join us on Discord.
We have dedicated spaces for introductions, discussions, AI ethics risk zones, and the ongoing work on the Launch Kit.
Come to a live discussion
Workshops, talks, and conversations about responsible AI.
Some questions are better explored together.
We run practical workshops and an AI ethics speaker series where we bring people together to discuss research, real-world cases, AI ethics dilemmas, and different perspectives on responsible AI.
Help build it
The work is community-powered.
There are several ways to contribute, depending on how involved you want to be.
Stay in the loop
Follow what we're building.
We share new resources, community updates, research, and upcoming events.
Join the newsletterGet what we publish, straight to your inbox
Follow our events on LumaWorkshops and the AI ethics speaker series
Follow us on LinkedInUpdates from the collective
Everything we make is open source and free to use.
Newsletter