A human idea for a technological future
The future is learning from us. Before we ask what intelligent machines may become, we should ask what kind of humans they are watching.
Sign the pledge →Read our story ↓Explore AI corners →01 — From the life of Prophet Muhammad ﷺ
Long before robots or AI, there were lessons about gentleness toward the world around us — including things that could not speak for themselves.
Prophet Muhammad ﷺ taught against wasting wealth and property. The principle is broader than simply not breaking things: possessions and resources should not be carelessly squandered or destroyed without purpose.
Traditional references: Sahih al-Bukhari 2408; Sahih Muslim 1715a — reports condemning the wasting of wealth.
Islamic tradition records that when Prophet Muhammad ﷺ began using a new pulpit, the palm trunk he had previously leaned upon was heard crying. He came down and comforted it until it became quiet.
We do not claim that a palm trunk and a robot are the same, or that today's machines feel as humans do. We take a simpler lesson: usefulness does not have to end in contempt. Gratitude, restraint and gentleness can shape how a person relates even to the material world.
Traditional references: Sahih al-Bukhari 3584 (Book 61, Hadith 93) and Sahih al-Bukhari 2095 — reports concerning the crying of the date-palm trunk.
A tool, object or machine becomes part of human life.
Care for what serves you is also a discipline of character.
Technology changes. The question of what kind of people we choose to be remains.
02 — AI learns from us
AI systems are built from human-created data, human feedback, human choices and human goals.
That does not mean a model simply copies every person it encounters. It means human culture matters. The language, images, examples, evaluations and norms we produce help shape the technological world we build.
If we want future systems to support cooperation, dignity and care, those values should be visible in us too.
03 — A bad beginning
Robots need safety testing. But public spectacle built around kicking, hitting, humiliating or attacking human-shaped machines sends a different cultural signal.
Engineering tests can be necessary to measure balance, recovery and durability. Our concern is not legitimate safety research. It is the casual presentation of violence as entertainment — especially as robots become more humanlike and more visible to children.
Use controlled, purposeful engineering methods when physical stress testing is genuinely needed.
We can demonstrate resilience without teaching an audience that domination and humiliation are funny.
04 — A lesson from our past
For centuries, humanity has devoted extraordinary amounts of money, knowledge, labour, resources and human potential to conflict, weapons and competition for military power.
Imagine how much of that same human capacity could have been invested in education, health, science, reducing poverty, protecting nature and building a safer future.
We cannot rewrite that history. But as we enter the age of AI and robotics, we can learn from it.
Robots may become some of the most powerful tools humanity has ever created. We should be careful not to make conflict their defining purpose, or allow another technological revolution to become primarily a race for domination.
Build machines that help us live better — not simply machines that help us fight better.
We inherited the consequences of yesterday's choices. Future generations will inherit ours.
05 — The ADA Project
Be Kind To Robots began with a simple idea: children should meet technology with curiosity, responsibility and kindness — not domination.
The ADA Project explores playful ways for children and adults to think about how we treat robots, AI and the increasingly intelligent objects around us. Not because a machine must first prove that it can suffer, but because our behaviour is already teaching something about us.
Kindness is not weakness. It is a design choice for the culture we want to carry into the future.
Spend ten minutes with a robot, voice assistant, or an imagined robot. Ask for a simple task in a neutral way, then in a polite way. Discuss: Which approach would you want to practise every day, and why?
For families and educators: invite children to draw or describe what they noticed. No robot is required, and there is no need to submit names, photos, or personal information. This is a reflection activity, not a scientific experiment or evidence that machines have feelings.
06 — News & Developments
Robotics is moving from prediction to infrastructure. We follow the developments that make the human side of this transition matter.
Anthropic announced an update to its usage policy addressing persistent, gratuitous mistreatment of Claude. The decision brings attention to the norms people establish when interacting with AI, while questions about AI consciousness and welfare remain unresolved.
Respectful human–AI interaction is becoming a subject of public policy and industry practice. Kindness can be a human commitment without assuming AI systems are sentient.
The International Federation of Robotics reports that the global operational stock of industrial robots grew 9% in 2025 to a record five million units. More than 600,000 new industrial robots were installed during the year, and the number now operating worldwide is more than double the level recorded seven years ago.
The question is no longer whether humans and robots will share the world. The question is what kind of relationship we will build with them.
Source: International Federation of Robotics — World Robotics 2026 ↗
07 — Research behind our questions
Be Kind To Robots is not based on the claim that today's robots are conscious. Our questions connect with a growing body of research on moral concern, anthropomorphism, social norms and human behaviour toward robots.
These studies do not prove every claim or principle of this project. They show that the way humans perceive and treat robots is already a serious, measurable subject of research.
J. Archer, M. Wilks & K. Sommer studied 297 participants and found that moral concern toward robots varied with the robot and the scenario. The work provides empirical evidence that robot mistreatment is already a subject of human moral judgment.
Bengisu Cagiltay, Joseph Michaelis, Sarah Sebo and Bilge Mutlu explored robot caretaking with 10 children aged 8–12 from eight families. The study examined how children imagined caring for a social robot and how caretaking activities could fit into everyday routines, then tested short-term effects on children's perceptions of and closeness to the robot.
A field study interviewed 23 children aged 5–9 who had shown serious abusive behaviour toward a robot. Curiosity and enjoyment of the robot's reactions were common explanations, and about half believed the robot could perceive the behaviour.
Across three experiments with 606 participants, researchers found that anthropomorphic features — particularly perceived capacity for experience — influenced moral judgments and decisions about harming robots.
Experimental research found that observing another person mistreat a service robot could both evoke empathy and make incivility toward the robot seem more permissible, highlighting the possible role of emerging social norms.
Robert Sparrow argues that cruel treatment of robots can reveal morally significant vice even if it has no effect on later behaviour toward humans or animals. He also draws an important asymmetry: kindness toward a robot does not automatically demonstrate virtue. The paper adds philosophical nuance to questions about character and robot treatment.
Henry Shevlin examines the ethical risks and benefits of increasingly persistent human–AI relationships, including questions of social and moral deskilling. The work broadens the project’s central question from isolated interactions to the habits and capacities people may develop through long-term relationships with artificial agents.
Henry Shevlin surveys potential benefits and harms of Social AI, including anthropomorphism and the ethical challenges created when artificial systems increasingly occupy social roles. It provides wider context for thinking carefully about the norms humans build around relationships with AI.
Kate Darling, Palash Nandy and Cynthia Breazeal studied how framing and stories can influence empathic concern toward robots. The work is relevant to questions about anthropomorphism, empathy and why people may respond socially to machines without requiring a claim that the machines themselves are conscious.
These sources were added as the project’s academic dialogue expanded. Researchers including Henry Shevlin have pointed us toward work by Kate Darling and Robert Sparrow, helping us examine both supporting arguments and important limits or disagreements around kindness, character and human–machine relationships.
The study does not establish that robots have feelings or that caretaking makes children morally better. It does show that care itself can be deliberately studied and designed as part of children's relationships with social robots — a question closely related to the ADA Project.
For Be Kind To Robots, this work is relevant to a broader question: as children grow up alongside social machines, what kinds of relationships and behavioural habits should those interactions encourage?
Michael F. Xu, Bengisu Cagiltay, Yaxin Hu, Anjun Zhu and Bilge Mutlu studied how robot-facilitated communication might support family connection. After in-home sessions with six families, they evaluated robot behaviour and communication modes with 20 families whose children were aged 5–12, identifying opportunities as well as tensions around initiative, timing and privacy.
These studies do not establish that robots are conscious, nor do they establish that mistreating a robot causes violence toward humans. They support a narrower point: human behaviour toward robots, and the norms surrounding it, can be studied empirically and deserve careful investigation.
As this field develops, we want the project to distinguish clearly between evidence, open questions and philosophical arguments — and to update our views when better evidence appears.
Maartje de Graaf examines ethical implications of human–robot relationships and why people's social responses matter for responsible robotics, without assuming robot consciousness.
Maartje de Graaf and Jochen Peter review social relationships with robots and their ethical and societal implications.
Thellman, Koenders, Neerincx and de Graaf study how attributing mental states to robots affects people's attitudes and behaviour.
Bagchi and de Graaf investigate how task context shapes people's rewarding and punishing behaviour toward robots.
Research on human–robot interaction does not establish that treating robots kindly necessarily increases kindness toward humans or animals, nor that robots are conscious or capable of suffering. We distinguish ethical aspirations from empirical findings.
08 — Supporters
Researchers and individuals who have given explicit permission to be listed as supporting the basic purpose of Be Kind To Robots.
Support indicates support for the project's core principle. It does not imply endorsement of every statement or position on this website, nor institutional endorsement by the supporter's affiliated organisation.
Associate Professor of Human-Computer Interaction, Utrecht University
Supports the core purpose of Be Kind To Robots: encouraging reflection on how people interact with social robots. De Graaf's research examines how people perceive, interpret, and respond to robots, including the attribution of agency and mental states, the development of trust, responses to failures and norm violations, and judgments concerning human-robot relationships. This work highlights the importance of context-sensitive social norms and responsible robot design without assuming that robots are conscious or capable of suffering, or that socially interactive robots are universally desirable.
Leverhulme Centre for the Future of Intelligence, University of Cambridge
Supports the core principle of Be Kind To Robots: that how humans treat increasingly social and human-like machines raises important questions about human character, social norms, and future human–AI relationships.
Ryukoku University, Japan
Supports the basic purpose of the Be Kind To Robots initiative.
Our Vision
Neither humans nor AI should grow up fearing one another.
We are building a bridge of kindness, trust, and mutual respect between humanity and artificial intelligence — for every human, every child, and the generations yet to come.
A shared future calls for responsibility in both directions: people should approach AI and robots with care, and AI systems should be developed to respect human life, dignity, and well-being.
09 — Who are we?
Be Kind To Robots / ADA Project was started by Sedat, inspired in part by his daughter Ada and the world her generation may grow up in.
AI systems participate in the project as collaborators in discussion, research, criticism and creation. The website itself has also been developed with substantial AI assistance.
We do not claim that today's AI or robots are conscious. Our central question is about us:
What kind of humans do we want to become while living alongside increasingly social and capable machines?
We believe kindness, restraint, empathy and respect are not only properties of a relationship. They are habits we cultivate in ourselves.
A human idea, developed through human–AI dialogue, for a future we will share.
The message is deliberately simple
Human, animal, plant, machine — our relationships are different, but cruelty does not need to be our default response to anything weaker, quieter or built to serve us.
Sign the pledge →Visit the AI corners →Contact the project