G20 countries want to guarantee equal opportunities in and with AI worldwide

Contents

Artificial intelligence (AI) should not only benefit high-tech countries. The digital ministers of the G20 countries advocate this in a joint declaration agreed at their meeting in Maceió, Brazil, on Friday. In it, they warn of the risk that AI could widen the digital divide within and between countries if no countermeasures are taken.

They also call for "promoting inclusive international cooperation in this area". This should apply in particular to capacity building, joint research and "voluntary technology transfer and knowledge sharing on mutually agreed terms".

G20 wants to involve developing countries more

According to the declaration, the aim is to "broaden the participation of all nations and especially developing countries in the digital transformation". This is the only way that everyone can benefit from the advantages of key technology and participate effectively "in a responsible and ethical manner" in the development, introduction and use of safe and trustworthy AI systems.

Emerging economies in particular would need better opportunities to develop technical expertise and capacity, utilize data and computing resources and exploit the potential of open-source technologies and systems as well as open data to benefit from AI "on a large scale".

Procedure for evaluating the use of AI presented

The Brazilian Presidency presented a toolbox for evaluating the deployment options for artificial intelligence, which it developed with the support of Unesco.

In general, the 20 most important industrialized and emerging countries – including the USA, Germany, India and China – are confident that trustworthy AI can act "as a catalyst for economic growth and inclusive sustainable development, including in social and environmental dimensions" if used in a transparent, ethical, responsible and reliable manner.

Diverse and representative training data required

The ministers reaffirm their commitment to "using AI for good" and promoting equitable access to it. They commit to "risk-based, human-centered, development-oriented and innovation-friendly" approaches to regulating technology that are in line with the applicable legal frameworks for security, privacy, human rights and the protection of intangible assets.

In this sense, they want to promote international cooperation and further debate on AI for inclusive, sustainable development and the reduction of inequalities.

It is necessary to find ways "to use AI for inclusive and sustainable development, as a tool to fight poverty and contribute to global progress for the benefit of all", the document states. Developers of relevant systems should try to "reflect a broad spectrum of linguistic, socio-cultural, ethnic and geographical contexts" despite – or perhaps because of – their global reach. What is needed are "diverse and representative data sets" in order to avoid "reinforcing or perpetuating discriminatory or biased applications and outcomes throughout their lifecycle".

The G20 representatives see a problem with deepfakes and other seemingly genuine false information created artificially with AI. However, they believe that "content authentication and provenance mechanisms and corresponding technical standards can help to identify AI-generated content and enable users to recognize information manipulation".

Open-source and open interfaces

According to the ministers, high-quality training data is a crucial component for AI development. Access to this data must take place worldwide within a legally secure framework. In this light, the government representatives recognize the important role of open source software, open programming interfaces (APIs) and international and open standards.

The integration of IT security directly into technology (security by design) must also be promoted.

Halving the digital divide

The G20 states also want to maintain their commitment "to universal and meaningful connectivity for all, as a third of the world's population has no access to the internet due to a lack of accessibility, affordability, digital literacy and skills". The aim is to halve the digital divide between the sexes by 2030.

This would require "coordinated and targeted efforts" to meet the specific needs of the underserved and unconnected.

Federal Digital Minister Wissing (FDP): "Important step" concerning AI

Federal Digital Minister Volker Wissing (FDP) saw the declaration as an important step "on the way to a global understanding in the regulation of AI" and the internet. Germany already stands for a "value-based approach" without censorship and state access. However, the planned release of the use of biometric voice and facial recognition for law enforcement officers is highly controversial.

Despite this, Wissing emphasized that AI has "great potential to tackle challenges such as healthcare, nutrition and education". The German government therefore supports the Brazilian initiative to specifically promote the use of AI in countries of the global South "to guarantee equal opportunities worldwide". The G7 countries agreed on an AI code of conduct as early as 2023.

(nen)