From competition to collaborations: CAIGP and the future of AI

As the world grapples with challenges and opportunities presented by AI, the Collaborative AI for Global Partnership (CAIGP) is expected to play an increasingly important role in the years to come. CAIGP focuses on fostering international collaboration and partnerships to ensure that the development and use of AI is equitable and beneficial for all.

With global partnerships for AI in focus, the three-day Global Partnership on AI (GPAI) Summit 2023 is being held in Delhi. The mega-event is marked by the participation of delegations of 28 member countries and European Union, including five Ministerial delegations (Canada, France, Japan, Turkiye, UK), and 67 GPAI experts and more than 120 industry leaders.

At the GPAI Summit 2023, the session on Collaborative AI for Global Partnership (CAIGP) - Global Cooperation for Equitable AI (Public) aimed at promoting faster innovation as countries working together share resources and expertise, thereby accelerating the development of AI technologies. It further emphasized on ensuring that the development and deployment of AI adhere to ethical standards that prioritize the protection of human rights, transparency, and accountability, while also avoiding any potential negative impact on individuals or communities.

Artificial General Intelligence (AGI) has the potential to resolve some of the world's most pressing challenges, such as climate change, poverty, and disease. However, it is important to ensure that the benefits of AGI are shared by all and do not exacerbate existing inequalities.

Speaking on AGI, Ms. Anna Makanju, Vice President, Global Affairs at OpenAI, in her keynote address, said: “Young entrepreneurs and civil society leaders are clearly going to be the foundation of India’s future in AI. As other countries work to create the educational building blocks to ensure that their workforces will succeed in AI, they should look to India, which has been successful for so long at developing top-tier technology talent.”

Makanju said that her company’s goal is that artificial general intelligence (AGI) benefits everyone. “Our charter lays out four principles to anchor this work – first, we commit to working towards broad distribution of the benefits of AGI; second, we commit to maximizing the long-term safety of AI systems; third, we commit to cooperating with other research and policy institutions; and fourth, we strive to be on the cutting edge of AI capabilities serving as a technical leader in the space.”

Further, Dr. John Ashley, Director, NVIDIA AI Technology Centers, shared his perspective about the inclusivity of the AI ecosystem and challenges to the hardware advancements. “One of the things that NVIDIA has been focusing on for a long time is what we call ‘democratizing AI’. We want to make the software, the hardware, the solutions, the systems available to as many people as possible,” Dr Ashley said.

Pointing out the ‘AI divide’ between Global North and Global South, Mr Ashley called for reduced barrier entries in the latter so that it’s easier to fine-tune AI models and adopt them. “There are countries in the Global South that have the education systems and the resources to deliver the talent. The challenge for them becomes the data.”

Meanwhile, Mr. Harsh Dhand, Head, APAC Google Research and AI Partnerships, Google, shed light on ways to bridge the gap between leading AI economies and the Global South, with respect to the software and algorithmic perspective.

“We at APAC make sure that we have the data generated, collected, curated across the region, with India being the epicentre, to ensure that the data is equitable in terms of quantity and corpus. Whether it is agriculture, health or education, we are making sure to collect the data which is trained in the region. So, we are ensuring that regional LLMs, not just for Google but for everyone – ecosystems, OpenAI, startups – to train using that data.”

Speaking about the importance of building guardrails for generative AI, Ms. Gabriela Ramos, Assistant Director-General (ADG) for Social and Human Sciences at the UNESCO, said that the “amazing transformative power of AI was still highly concentrated in few companies and countries”.

“There has been a 28-fold increase in investments in AI over the last decade, but is has been more in US, China and some other parts of the world. So, it’s highly concentrated in terms of development. We have to tackle this question because eventually this is what leads to high concentrations at the top and discriminatory patterns at the bottom. Not only should we explore the possibilities of Global South to take advantage of the technology, but it should be able to develop its own systems.”

Echoing similar sentiments, Mr. Vilas Dhar, President and Trustee at the Patrick J. McGovern Foundation, said that “while we talk about the incredible possibilities that AI has to offer, we continue to struggle with the foundational question – What it means when you take a conversation about the possibility of technology and evaluate it through the lens of equality and justice.”

“I am so excited for India’s leadership at GPAI over the next year because in many ways, I think the question that we face now is how do we take policy conversations and turn them into implementation conversations,” he added.

The session concluded discussing both challenges and opportunities for global cooperation on equitable AI. Countries may have different priorities for AI development and use, which can make it difficult to reach agreements on global standards. Moreover, robust international agreements will be required to address the concerns about data privacy and security while sharing data across borders. Developing countries may lack the resources needed to develop and use AI technologies, which could exacerbate existing inequalities, as in the case of ‘AI divide’ between Global North and Global South. Moreover, international cooperation and dialogue will be imperative to address ethical concerns surrounding the development and use of AI, such as bias, discrimination, and the potential for misuse.