EU guidelines on ethics in artificial intelligence: Context and implementation
11
conduct in the field of AI. Preparatory work for establishing an AI ethical framework is also ongoing
in India, New Zealand, Singapore, South Korea and Japan.
30
Some international organisations
are engaging in setting international rules in the field of ethics and AI. In May 2019, the OECD and
associated nations adopted a non-binding list of guidelines
for the development and use of AI.
The AlgorithmWatch AI Ethics Guidelines Global Inventory lists all of the ethical frameworks and principles
being developed across the globe. Recent years have seen a flurry of initiatives from companies, governments,
NGOs and research bodies to propose ethical rules on AI. The ethical principles laid down in other jurisdictions
seem relatively similar to those of the EU (though less detailed) and are essentially of a self-regulatory nature,
even though there is growing demand for more government oversight.
Outlook
Policy-makers all over the globe are looking at how to tackle the risks associated with the
development of AI. In April 2019, the EU published its guidelines on ethics in AI, becoming a front-
runner in the setting up of a framework for AI. Ethical rules on AI, where such exist, are so far
essentially of a self-regulatory nature, and there is growing demand for more government
oversight. In the EU, there are strong calls for clarifying the EU guidelines, fostering the adoption of
ethical standards and adopting legally biding instruments in order to, inter alia, set common rules
on transparency, set common requirements for fundamental rights impact assessments and
provide an adequate legal framework for face recognition technology.
MAIN REFERENCES
Boucher P., How artificial intelligence works, briefing, EPRS, March 2019.
The Age of Artificial Intelligence, European Political Strategy Centre, European Commission, March 2018.
Ethics Guidelines for Trustworthy AI, Independent High-Level Expert Group on Artificial Intelligence,
European Commission, April 2019.
Artificial Intelligence: A European Perspective, JRC, European Commission, 2018.
Kritikos M., Artificial Intelligence ante portas: Legal & ethical reflections, briefing, EPRS, March 2019.
A governance framework for algorithmic accountability and transparency, study, Scientific Foresight Unit
(STOA), EPRS, April 2019.
Understanding algorithmic decision-making: Opportunities and challenges, study, STOA, EPRS,
March 2019.
ENDNOTES
1
See definition provided by the High-Level Expert Group on Artificial Intelligence (glossary section of the Ethics
Guidelines for Trustworthy AI). There is no commonly agreed definition for AI. For an overview of the notion of AI and
the difficulty in defining it, see Philip Boucher's 2019 EPRS briefings on How artificial intelligence works and on Why
artificial intelligence matters.
2
See EPRS briefing on Economic impacts of artificial intelligence by Marcin Szczepański, July 2019.
3
See Data quality and artificial intelligence – mitigating bias and error to protect fundamental rights, FRA Focus, EU
Agency for Fundamental Rights, June 2019.
4
See two EPRS publications by Mihalis Kritikos: What if algorithms could abide by ethical principles?, ‘at a glance’ note,
November 2018; and Artificial Intelligence ante portas: Legal & ethical reflections, briefing, March 2019.
5
Furthermore, the Parliament adopted an own-initiative-resolution on a Comprehensive European industrial policy on
artificial intelligence and robotics in February 2019. The European Economic and Social Committee issued an own-
initiative opinion on AI in May 2017 also calling for a code of ethics for the development, application and use of AI.
6
Furthermore, on 10 April 2018, 25 European countries signed a Declaration of cooperation on Artificial Intelligence.
7
See definition provided by the High-Level Expert Group on Artificial Intelligence (glossary section of the Ethics
Guidelines for Trustworthy AI).
8
See The Age of Artificial Intelligence, European Political Strategy Centre, European Commission, March 2018.
9
Ibid. See as well Artificial Intelligence: A European Perspective, JRC, European Commission, 2018.
10
EU action could also be taken to clarify the issue of liability of damages and measures to encourage data-sharing.