The United Nations Educational, Scientific, and Cultural Organization (UNESCO) has published a global standard on artificial intelligence (AI) ethics that almost 200 member nations are expected to embrace.
The standard establishes shared values and concepts that will guide the creation of the legal infrastructure required to support AI’s healthy growth. According to UNESCO, there are various advantages to using AI, but there are also drawbacks, such as gender and ethnic bias, substantial risks to privacy, dignity, and agency, dangers of mass surveillance, and greater use of inaccurate AI technologies in law enforcement.
This is not the first time the concerns about the accountability of leveraging AI technologies have come into the limelight. While sure, it ushered discussions on the ethics of AI and a regulatory check on the possibility of misuse of AI outstripping its benefits and promises; there is still a gap.
While AI technologies offer enormous promise for social and economic growth, policymakers face complicated and divergent hurdles in implementing them. Bias, stereotyping, and prejudice are among issues that AI brings to the table when such discussion happens. Amid these concerns, AI-generated analysis is increasingly being used to make decisions in both the public and private sectors. Hence, UNESCO has asked for AI to be created in a way that promotes fair outcomes.
Presenting the 28-page document, formally titled “Recommendation on the Ethics of Artificial Intelligence,” Audrey Azoulay, director-general of the UNESCO, stated that the content of the recommendation focuses on the following pointers:
- Individuals will be better protected by ensuring openness, agency, and control over their personal data, which will go beyond what digital companies and governments are doing. It emphasizes that everyone should be allowed to access and even delete their personal data records.
- The standard strictly prohibits the use of artificial intelligence (AI) systems for social scoring and mass monitoring. It insists that while building regulatory frameworks, member states should keep in mind that ultimate responsibility and accountability must always rest with people and that artificial intelligence technology should not be given legal authority in itself.
- The ethical impact assessment is designed to assist nations and businesses in assessing the impact of AI systems on persons, society, and the environment as they develop and deploy AI systems.
- The standard suggests that governments evaluate the AI system’s direct and indirect environmental impacts throughout its life cycle, including its carbon footprint, energy consumption, and the environmental effect of raw material extraction for manufacturing AI technologies.
The Recommendation will include provisions to prevent real-world biases from being repeated online, as well as tangible policy actions based on universal values and principles. It will also instruct UNESCO to assess each country’s progress in the field of AI to assist them in the implementation phase.
In 2018, Audrey Azoulay, the Director-General of UNESCO, announced an ambitious initiative of establishing an ethical framework for using artificial intelligence across the world. Three years later, owing to the mobilization of hundreds of experts worldwide and extensive international talks, the 193 UNESCO member nations have recently officially embraced this ethics of AI standard framework.
According to Azoulay, “The world needs rules for artificial intelligence to benefit humanity. The recommendation on the ethics of AI is a major answer. It sets the first global normative framework while giving states the responsibility to apply it at their level.” Azoulay asserts that UNESCO will support its 193 member states in its implementation and ask them to report regularly on their progress and practices.
Meta (formerly known as Facebook) has been at the center of various controversies surrounding the ethics of AI in recent years. Cambridge Analytica, a now-defunct British political consulting business, exploited Facebook’s data to sway the Brexit referendum in the United Kingdom and Donald Trump’s victory in the United States. In 2018, Timnit Gebru, a former employee of Google’s Ethical AI team, and Joy Buolamwini, a researcher, demonstrated that face recognition software was less accurate in recognizing women and persons of color than it was in identifying white men.
Unesco’s proposals come less than a month after China released its own set of AI ethical principles, focusing on user rights and aligning with its ambition of becoming a global AI leader by 2030.
Meanwhile, AI experts are concerned that a few African voices have been included in the worldwide ethical regulations that offer guidelines for AI research. This is crucial as African countries are investing more in AI and machine learning research and development today. Data Science Africa, Data Science Nigeria, and the Deep Learning Indaba with its satellite IndabaX events, which have taken place in 27 African nations to date, demonstrate the interest and public investment in the AI disciplines.
While, surprisingly, China came out in favor of the 28-page document suggestion as one of UNESCO’s 193 member nations, surveillance cameras with biometric facial recognition dot the cityscape of the Mandarin nation. China is already in hot waters for its role in technologically aided persecution of the Muslim Uyghur minority in the autonomous area of Xinjiang, as well as the struggle against Hong Kong’s democracy movement — fueling the fears of the use of AI to control public behavior.
Hundreds of towns, from Dubai to Nairobi, Moscow to Detroit, have placed cameras with FRT, with the assurance of feeding data to central command centers under smart city crime solutions. And it’s worth noting that most of these cities’ governments didn’t get public approval for the unbridled use of face recognition in the name of law enforcement.
On the brighter side, at the same time, several countries, such as Belgium and Luxembourg, have stated their opposition to the use of face recognition technology.
The success of this initiative lies with governments around the world, willing to adhere to the basic protocols and guidelines that will ensure ethical use of AI in the long run. A comprehensive follow-up will be necessary to ensure that the UNESCO recommendation can successfully offer universal standards for policy and legislation. As per the document, governments should create a regulatory framework that lays out a mechanism for conducting ethical impact assessments on AI systems to foresee outcomes, minimize risks, avoid adverse repercussions, increase citizen engagement, and address societal concerns. Algorithms, data, and design processes should all be auditable, traceable, and explainable.
While this is not a one-day assignment, experts hope this will encourage more conversations. Other international bodies have been working on AI ethics as well. For instance, the Organization for Economic Cooperation and Development (OECD) issued “Principles on Artificial Intelligence” in 2019, which supports “respect [for] human rights and democratic ideals” while adopting the technology.