The Data Oath
The Data Oath
Welcome to The Data Oath
The Data Oath was founded by Christopher Sanchez and his team to help clients, companies, and governments understand the goals and vision when designing intelligent systems. He believed that the problem in AI Ethics isn’t setting the bar too high, but setting it too low because of how complicated the topic can be.
If you would like to use The Data Oath in your company you are free to do so, without charge, or formalities. We just ask that you link back to this site and if you would be so kind as to shoot us a message letting us know you liked it enough to use.
If you would like to learn more about how AI Ethics would benefit your organization, reach out to us, we’d love to hear from you.
Organizations must ensure that all data and systems are treating all users equally in regards to outcomes, scores, recommendations, etc. There shouldn’t be any group or groups that are benefiting more than other groups. Nor should there be any group or groups that are adversely impacted compared to others. The data and systems should be regularly reviewed and verified from a design, coding, and implementation standpoint.
Organizations must ensure that all system recommendations, scores, results, etc. are explainable and transparent to a non-technical user. Systems should be designed to provide an explanation road map for how it arrived at any decision when dealing with human users.
Organizations must ensure the intelligent systems they design, use, and deploy improve the quality of life for all users when compared to a human performing the same task in a material way. Organizations should not deploy systems that degrade the quality of life for any of their users in any material way. The Quality of Life factor of every system should be regularly reviewed from a design, data, coding, and outcome standpoint.
Organizations must ensure that all raw data, system logic, and weightings are available to users upon request for review, to be updated or contested. Users should have unfettered access to all data any institution uses when scoring a user for any purpose. No institution should have access to data that the user doesn’t also have access to. Users should be able to review, update, contest, and delete any information with any institution.
Sometimes the best way to understand the importance of something is through a story over a meal. So let us enjoy a virtual dinner where an AI system isn’t so ethical with your dining experience.
Let’s start with a short story about the importance of Ethical AI:
Data should be taken seriously.
Ethics are a work in progress, pushed forward by the changes in society and the standards society wishes to obtain. At the moment, the world is focused on defining the ethics that will govern artificial intelligence that in many ways will and does govern all of us. Artificial intelligence is playing a larger role than ever in all of our lives. From which of us get into college, to landing the right job, getting credit, or how we are sentenced in court. There will be many organizations, governments, and individuals suggesting the right way to move forward. We believe that we should participate in the discussion to help move the conversation in the right direction.
At The Data Oath, we take the position of common sense and rational expectation when designing the framework for AI Ethics. We don’t seek to propose standards that no company, government, or institution can uphold, but rather a set of Five Foundations that focus on nudging technological ethics in the right direction. The Five Foundations have been designed to set the floor for Ethical AI when dealing with Human Users. These foundations are meant to be built upon within the spirit and intention they were proposed. Over time, the methods used to achieve Ethical AI will change, but the foundations through which to envision those methods will remain the same.
Below we have listed the Five Foundations for Ethical AI. These are the Five Foundations that any company can use to evaluate what they are doing, how they are doing it, and if the actions being taken are in alignment with the spirit of the Five Foundations.
Transparent + Explainable
Improves Quality of Life
Accessible + Contestable
Organizations must actively look for ways to include more people, groups, races, genders, orientations, and other crucial data points to represent better the users and communities the systems are meant to serve. The systems and data should be regularly reviewed, verified, and updated from a design, coding, and implementation standpoint as needed.