The Mindful Automation Manifesto

Automation has already become an integral part of our lives, and is set to embed itself in our lives even more intimately. In more ways than we might imagine, we already rely on automation every day; people using increasingly sophisticated tools is what allows our civilization to exist in the first place.

At the heart of this revolution lies the promise of increased efficiency, productivity, and innovation. However, as we embrace these technological advancements, we must not lose sight of what makes us uniquely human – our ability to think critically, empathize, and act ethically. We must also remember to maintain a role for us, and allow people to find their purpose and meaning from as wide a variety of options as possible.

Rejecting the extreme views that we are facing either a Cornucopian utopia of productivity heralded by technological advancements, or a dystopian hell where humanity is over-run by its own creations guided by the worst traits of humanity, the Mindful Automation movement seeks to create a more harmonious and sustainable future; one where humans and machines work together, but in very different terms than that collaboration is often implemented today. The future we should strive for is one that integrates technology, including AI, into our world in a way that upholds the highest ethical standards and fosters human growth and happiness.

Our manifesto is built upon the following principles. Conscious of the often-times vague nature of principles, there are some possible practical steps outlined for each of the principles. They are meant to be illustrative, as prescriptive guidance that would apply in all situations is neither feasible nor warranted.

Principles Practical Steps (examples)
1. Ethical Decision Making: Automation should not absolve us of responsibility or accountability. We must ensure that the ethical implications of automated systems are carefully considered and that moral values are embedded throughout their development and deployment. We must consider the ethical implications of automated systems beyond the plethora of Ethical AI frameworks, embedding moral values in their development and deployment, and closing the gap between ethical AI frameworks and their implementation.Create an ethics committee made up of diverse stakeholders who can review and provide input on automation projects. Establish protocols that ensure ethical considerations are taken into account at every stage of an automation project, from initial conception to deployment and maintenance​.
2. Human Wellbeing-Centric Design: Human Factors Engineering is necessary, but not sufficient. Prioritize human well-being holistically, considering physical, mental, emotional, and social needs, while acknowledging the unique capabilities of machines. Balance relinquishing certain responsibilities to automation with the need to support skill maintenance and development, and prevent humans from becoming the “moral crumple zone” in any system. Any implementation of automation should also ensure the core components of human motivation – autonomy, mastery, and purpose – are not negatively impacted.Involve end-users in the design and testing process to ensure that automation tools are developed with their needs in mind. Regularly survey users on their experiences with automated systems and adjust as necessary to ensure their well-being is documented and prioritized​.
Resilience and Adaptability: Mindful automation recognizes that tightly coupled and complex systems are inherently unpredictable and dangerous. Wherever possible, tightly coupled and complex systems must be avoided; where unavoidable, they require specific procedure, the right culture, and strong governance to ensure the safety of the system. We must build resilience into our automated systems on a holistic level, allowing systems to withstand and learn from disruptions, while also ensuring that humans retain the skills needed to intervene, adapt and take over when necessary.Design automated systems with fallback options that allow for human intervention when necessary. Conduct regular resilience testing and adapt systems based on the outcomes of these tests​.
Collaboration, Inclusivity, and Global Perspective: The development and implementation of mindful automation must involve collaboration between diverse stakeholders from around the world, including technologists, ethicists, policy-makers, and others, as well as diversity within each stakeholder group and in all levels of the system. By fostering inclusive dialogue, we can create systems that are more equitable and beneficial to all members of society.Establish a diverse project team, including representation from different cultures, genders, ages, and professional backgrounds. Include the voices of all stakeholders in decision-making processes through regular meetings and open channels of communication.
Lifelong Learning and Skill Development: As automation continues to transform the workforce, we must invest in continuous learning and skill development for individuals to thrive in the changing landscape. Where the required skills mix is changing, this must include supporting education and training programs that enable people to work effectively and safely alongside automated systems and harness their potential for personal and societal growth. Importantly, people potentially impacted by automation need assurances of continued livelihood and details of the available (re-)training programs before any implementation project can proceed to enable an informed “buy-in”. Furthermore, organizations should implement a Just Culture to enable effectively learning from any close calls, incidents and accidents without blame.Offer regular training and professional development opportunities for employees to learn about new technologies and automation tools. Have clear communication and support plans in place for employees who might be impacted by automation, including retraining and job transition programs, financed appropriately so there are no material negative impacts.
Transparency and Trust: For automation to be successful, it must be rooted in transparency and trust. This involves openly communicating the capabilities and limitations of automated systems, as well as providing opportunities for public engagement and input in their development and governance. People need to know when they are dealing with, or looking at output of, an automated system. Where possible (and obviously where legislated), people also need to be able to opt out of participating in an automated system and be offered an option to deal with a person instead, without being unnecessarily inconvenienced.Clearly communicate to all stakeholders about the capabilities and limitations of automated systems. Provide channels for feedback and make adjustments based on this feedback. Allow for opt-out options where possible and feasible.
Environmental Sustainability: Mindful automation acknowledges the need to protect our planet and its resources for future generations. We must strive to develop and deploy automated systems that minimize environmental impact and contribute to sustainable development goals. Analysis of the environmental impacts should include considerations of Jevon’s Paradox.Conduct environmental impact assessments for all automation projects. Strive to use technologies and methods that minimize environmental impact, such as energy-efficient hardware and algorithms.
Safety. We acknowledge no system, whatever its level of automation, can ever be 100% safe. We must ensure introduction of automation improves, or at a minimum, does not further compromise safety at the expense of speed, efficiency, or effectiveness. The threshold for considering a system “good enough” must be raised, for we know humans have a tendency to over-rely on automation, even when that reliance is ill-advised. This includes not only physical safety, but consideration of societal and other risks introduced by the systems.Perform regular safety audits and risk assessments of automated systems. Prioritize safety over speed or efficiency in the design and implementation of automation. Establish protocols for responding to safety incidents and learning from them to improve system safety.
Economic Considerations: We must recognize the economic impact of any type of automation on businesses and the workforce, and develop strategies to mitigate job displacement and facilitate transitions, including reskilling programs, job-sharing arrangements, and exploring the concept of universal basic income. At the same time, organizations implementing automation in any form must be able to do so in an economically viable manner.Carry out economic impact assessments before implementing automation projects. Develop strategies to mitigate job displacement, including reskilling programs. Make sure that the automation projects are economically viable and provide a return on investment​.
Regulation and Oversight: We acknowledge the role of governments and regulatory bodies in ensuring responsible, ethical automation, and will offer recommendations for creating and enforcing policies that adhere to the principles outlined in the manifesto.Comply with all relevant laws and regulations; not just the laws aimed at AI or automation. Engage with regulatory bodies and keep up-to-date with policy changes. Offer recommendations to policymakers based on the organization’s experiences with implementing automation.