Professor Marc Rotenberg, President of Electronic Privacy Information Center (EPIC), Member of Michael Dukakis Institute’s AIWS Standards and Practice Committee recently published the Universal Guidelines for AI. It will be opened for public sign-on and later released at the Public Voice event in Brussels on October 23, 2018.
The emergence of AI is transforming the world, steering science and industry to government administration and finance on a whole new direction. The rise of AI decision-making also implicates fundamental rights of fairness, accountability, and democracy. Many of them are unclear to users, leaving them unaware whether the decisions were accurate.
Aware of the current situation, Prof. Marc Rotenberg – President of Electronic Privacy Information Center (EPIC) proposed these Universal Guidelines to guide the design and use of AI, which serve as a part of the first layer of AIWS 7-layer Model built by Michael Dukakis Institute. The goals of this model are to create a society of AI for a better world and to ensure peace, security, and prosperity.
These guidelines should be incorporated into ethical standards, adopted in national law and international agreements, and built into the design of systems. The guidelines include twelve terms:
- Right to Transparency. All individuals have the right to know the basis of an AI decision that concerns them. This includes access to the factors, the logic, and techniques that produced the outcome.
- Right to Human Determination. All individuals have the right to a final determination made by a person.
- Identification Obligation. The true operator of an AI system must be made known to the public.
- Accountability Obligation. Institutions must be responsible for decisions made by an AI system.
- Fairness Obligation. Institutions must ensure that AI systems do not reflect bias or make impermissible discriminatory decisions.
- Accuracy, Reliability, Validity, and Replicability Obligations. Institutions must ensure the accuracy, reliability, validity, and replicability of decisions.
- Data Quality Obligation. Institutions must ensure data provenance, quality, and relevance for the data input into algorithms. Secondary uses of data collected for AI processing must not exceed the original purpose of collection.
- Public Safety Obligation. Institutions must assess the public safety risks that arise from the deployment of AI systems that direct or control physical devices.
- Cybersecurity Obligation. Institutions must secure AI systems against cybersecurity threats.
- Prohibition on Secret Profiling. No institution shall establish or maintain a secret profile on an individual.
- Prohibition on National Scoring. No national government shall establish or maintain a score on its citizens or residents
- Termination Obligation. An institution that has established an AI system has an affirmative obligation to terminate the system if it will lose control of the system
The Universal Guidelines for AI is currently open to public sign-on by individuals and organizations. On October 23, it will be introduced at the Public Voice event in Brussels.