This page collects resources grappling with the ethics of AI: laws, frameworks, literature, and controversies.
Laws or Proposed Laws
- Proposed Algorithmic Accountability Act of 2019, released 4/10/19. Basically would implement the GDPR’s DPIA process in the U.S. Summary.
Ethical AI Frameworks
- Future of Life Institute AI Principles, aka the Asilomar AI Principles, developed as part of the BAI 2017 conference in January 2017.
- Salesforce AI Principles, published January 16, 2018.
- Google AI Principles, published June 7, 2018.
- IBM’s Everyday Ethics for Artificial Intelligence, published September 5, 2018.
- The Public Voice’s Universal Guidelines for Artificial Intelligence, published October 2018. The Public Voice is an international coalition established in 1996 to promote public participation in decisions concerning the future of the internet.
- Microsoft AI Principles, published Dec. 17, 2018.
- European Commission Ethics Guidelines for Trustworthy AI, published April 8, 2019. Vague but laudatory set of principles published by an expert group of the European Commission. Discussion here.
- NYT New Work Summit Ethical AI Recommendations, published March 1, 2019.
- OECD Principles on AI, adopted May 22, 2019.
- Beijing AI Principles, published May 28, 2019. Discussion.
- Policy and investment recommendations for trustworthy Artificial Intelligence, published June 26, 2019. This is a follow-up report from the EU High-Level Expert Group on Artificial Intelligence.
Studies and Other Literature
- Discrimination through optimization, published 4/4/19.
- The IEEE’s Ethically Aligned Design.
- The Obama Administration’s Roadmap for AI Policy. News story.
Organizations and Programs:
- AI Now Institute. New York University interdisciplinary research center dedicated to understanding the social implications of artificial intelligence. Published a report on lack of diversity in AI researchers.
- Upturn.org. 501(c)(3) based in Washington, D.C. founded around Nov. 2017. Mission statement is to “promote equity and justice in the design, governance, and use of digital technology.” Two authors on Discrimination through optimization study.
- White House Executive Order on AI. Someone should really start thinking about this stuff? Discussion.
- OpenAI. Founded in 2015, they have a mission to ensure that artificial intelligence “benefits all humanity.” Microsoft invests $1B.
- The Center for the Governance of AI at the Future of Humanity Institute.
- 2019 July: Tacoma, WA convenience store uses facial recognition to deny entry to certain persons. Post. Meanwhile, ICE Used Facial Recognition to Mine State Driver’s License Databases and Detroit tracks residents.
- 2019 May: SF restricts agencies from using facial recognition technology. Post. See also using facial recognition in police investigations, plus a response.
- 2019 April: Amazon allegedly automatically tracks and fires employees who don’t meet performance criteria. And AI software that evaluates human call center performance.
- 2019 April: China facial tracking. China is surveilling its minority muslim population with AI. Post. Meanwhile, the U.S. rolls their own facial tracking. Post.
- 2019 March: Facebook housing suit. Facebook sued for housing discrimination in advertising; algorithms blamed. Discussion.
- 2019 March: Amazon facial recognition. Amazon’s technology is criticized as biased. Discussion. Meanwhile, Microsoft declines to sell their system on ethical grounds.
- 2019 February: OpenAI GPT-2. OpenAI publishes results of GPT-2 language generation model, but declines to release full model. Controversy ensues.
- 2018 April: Google Duplex. The demonstration appears to deceive humans by design. Article.