You can find more information on five points mentioned during Jovan Kurbalija's keynote address at the Glion’s human rights dialogue.
1. Bottom-up AI
Bottom-up AI challenges the dominant view that powerful AI platforms can be developed only by using big data, as is the case with ChatGPT, Bard, and other large language models (LLMs). Bottom-up AI is technically possible and ethically desirable. For more details on technical, ethical, and practical aspects of bottom-up AI, you can consult the blog text: To foster human freedom and prosperity, Artificial Intelligence must be developed bottom-up!
2. Rights to be imperfect
Right to be imperfect is essential for preserving our core humanity in the race with digital technology. You can read more on right to be imperfect.
3. SDGs as guardrails for AI developments
The global community is searching for guardrails for AI governance and development. You can read why 17 sustainable development goals (SDGs) are obvious and available solutions for AI guardrails.
4. Global Digital Compact and human rights
The UN Secretary General's Policy Brief 5 on Global Digital Compact has a high focus on human rights, especially in the sections on AI and data. You can find an analysis of the GDC's coverage around 'nothing new under the sun' (traditional language in this field) and new approaches/proposals.
Human rights are directly referenced in 3 out of 8 objectives with the ‘nothing new under the sun’ narrative:
- Objective C: Upholding Human Rights
- Objective F: Data protection and empowerment
- Objective G: Agile governance of AI and other emerging technologies
However, there are a few ‘new’ highlights and aspects of digital and human rights:
- The idea of a digital human rights advisory mechanism, to be facilitated by the Office of the UN High Commissioner for Human Rights and 'provide practical guidance on human rights and technology issues'. (Read more)
- An invite for countries to 'consider the adoption of a declaration on data rights' and to 'consider the call [...] to seek convergence on principles for data governance through a Global Data Compact'. (Read more)
- A suggestion to 'combine international guidance and norms, national regulatory frameworks and tech standards into a framework for agile governance of AI'. (Read more)
- The idea of establishing a high-level advisory body for AI to 'offer perspectives on how ethical, safety and other regulatory standards could be aligned, interoperable, and compliant with universal human rights and the rule of law'. (Read more)
5. Can AI create more convergences around human rights?
In the current geopolitical tensions, there is a lack of visibility of convergences in the regulation of data and AI among the main centres of digital governance: USA, EU, and China. For example, the EU’s GDPR and China’s PIPL data regulations are very similar. There are also many common points among the three key actors in AI governance.