UNESCO adopts recommendations on the Ethics of AI
UNESCO, in the 41st session of its General Conference, formally adopted a set of recommendations on the Ethics of Artificial Intelligence.
The first such exercise by a global standard-setting body, it took UNESCO two years to work out the recommendations. The process was carried out in two phases. The first involved creating a draft text by an Ad Hoc Expert Group, after consultation with a wide range of stakeholders from various fields. The second phase involved an intergovernmental process and negotiations on the draft text to produce a final version of the recommendations, which was adopted, on 24 November 2021.
UNESCO is the first such global instrument that covers all aspects of ethical considerations that are important for policymaking by member states as well as other private organizations influencing the development and application of AI.
UNESCO’s work on the Recommendation has built on the preliminary study on ethics of artificial intelligence of UNESCO’s World Commission on the Ethics of Scientific Knowledge and Technology (COMEST).
The objectives of this Recommendation (taken verbatim from the document) are:
(i) to provide a universal framework of values, principles, and actions to guide States in the formulation of their legislation, policies, or other instruments regarding AI, consistent with international law
(ii) to guide the actions of individuals, groups, communities, institutions, and private sector companies to ensure the embedding of ethics in all stages of the AI system life cycle.
(iii)to protect, promote and respect human rights and fundamental freedoms, human dignity and equality, including gender equality; to safeguard the interests of present and future generations; to preserve the environment, biodiversity, and ecosystems; and to respect cultural diversity in all stages of the AI system life cycle.
(iv) to foster multi-stakeholder, multidisciplinary and pluralistic dialogue and consensus-building about ethical issues relating to AI systems.
(v) to promote equitable access to developments and knowledge in the field of AI and the sharing of benefits, with particular attention to the needs and contributions of low to medium income countries including least developed countries, small island developing states, and landlocked developing countries.
The recommendations covered 11 policy areas for affirmative action for ensuring an ethical evolution of AI development and application. They are
· Ethical Impact Assessment
· Ethical Governance and Stewardship
· Data Policy
· Development and International Cooperation
· Environment and Ecosystem
· Gender
· Culture
· Education & Research
· Communication & Information
· Economy & Labour
The whole report containing detailed recommendations can be downloaded from the UNESCO site here.
Google launches GLaM: Generalist Language Model (GLaM)
GLaM allows for efficient scaling of Language Models with Mixture-of-ExpertsLink: https://arxiv.org/pdf/2112.06905.pdf
META’s BEAN Machine a PyTorch Based Probabilistic Model System
Bean Machine enables you to develop domain-specific probabilistic models, and automatically learn about unobserved properties of the model with automatic, uncertainty-aware learning algorithms. Compared to other machine learning approaches, probabilistic modeling offers numerous benefits such as Uncertainty Estimation, Expressivity, and interpretability
Distributed Artificial Intelligence Research Institute (DAIR)
Timnit Gebru launches AI research institute to spearhead community-rooted research
Timnit Gebru, who was terminated by Google as the technical co-lead of the latter’s Ethical Artificial Intelligence Team in December 2020, when she challenged Google management’s request to withdraw her unpublished paper on risks and biases of large language models, has launched Distributed Artificial Intelligence Research Institute (DAIR), that will ‘create space for independent, community-rooted AI research free from Big Tech’s pervasive influence.’
Link: https://www.dair-institute.org/press-release
Upcoming AI Conferences
ICML 2022 (Jul 17- 23, 2022)
Baltimore, United States
Link: https://icml.cc/Conferences/2022/
ECCV 2022 (October 24–28, 2022)
Tel-Aviv, Israel
Link: https://eccv2022.ecva.net/
AI Debate #3 (May 04, 2022)
by Montreal.ai and Gary Marcus
Link: https://montrealartificialintelligence.com/aidebate3/