Towards Comperehensive Approach to AI Design, Development, and Deployment: View from an international organisation

For instance, a policy chapter in the Recommendation concerns gender equality, stressing the need for gender diversity and inclusiveness in AI development teams, to mitigate bias and discrimination in AI systems and avoid exacerbating existing gender gaps. Additionally, the importance of assessing the direct and indirect environmental impact (e.g., CO2 emitted with AI, water lost, etc,) throughout the AI system's life cycle for member states and businesses was also emphasised. 

Concerns were also raised in relation to the need for AI governance for business enterprises and other stakeholders - both public and private, as e.g. larger companies have the resources to invest in AI technology while smaller ones struggle, thus leading to gaps with SMEs. An interesting development discussed was the use of agentic tools, a subset of AI development involving LLMs that can interact with other LLMs to mimic human expertise (e.g., expert marketing, expert in finance, etc.), presenting a complex picture of opportunities and challenges. While they may  contribute to democratise access to expert knowledge and boost productivity for SMEs, agentic tools and their use also raise concerns about possible job displacements, including for highly skilled professionals. As a result, UNESCO advocates for member states to support collaborative agreements with various stakeholders to bridge the skill gap through education and training programmes. 

It is worth highlighting that the UNESCO Recommendation on the Ethics of Artificial Intelligence demonstrates the importance of balancing global approaches with national and local needs. Policy decisions cannot be entirely isolated from international relations, investments and other external components, they are not mutually exclusive and overlap. Ensuring the development of culturally relevant AI approaches represents a complex but needed issue. While fostering regional collaboration and knowledge-sharing through UNESCO regional councils is a crucial step in capacity building, it is essential to critically reflect on the potential to coordinate approaches and governance. AI provides an opportunity for developing countries to address their specific challenges and despite potential infrastructure limitations, these nations are active participants in the global AI discourse. To truly empower developing nations, it is paramount to support the growth of local expertise and create an enabling environment for innovation that is grounded in local values and priorities, as well as global human rights.

Ultimately, governance and regulation for AI should not be viewed as a barrier to innovation, but rather as a tool to harness AI's potential for good. For example, regulation in the pharmaceutical industry did not hinder innovation; instead, it created a clear framework to ensure drug safety, efficacy, and public trust. Similarly, UNESCO’s Recommendation opens the discussion on how to fairly distribute the benefits of AI and making policy flexible enough to adapt to new developments. As AI continues to evolve, promoting international cooperation and advancing ethical guidelines are essential, as is equipping individuals and communities alike with the knowledge and skills necessary to navigate their lives and futures. 

Citations:

https://www.unesco.org/en/artificial-intelligence/recommendation-ethics?hub=32618

https://link.springer.com/chapter/10.1007/978-3-030-69128-8_15