Editorial 1: Democratising AI needs a radically different approach

Context

Not aiming to compete with or replicate the Big Tech model is essential; it should even look at championing ‘small AI’.

Introduction

The growing power and influence of Big Tech companies is a concern for policymakers worldwide. To break Big Tech’s hold over the Artificial Intelligence (AI) ecosystem and democratise AI development, India, like many other countries, is investing in sovereign cloud infrastructure, creating open data platforms and supporting local start-ups. However, these efforts are unlikely to be enough and may even deepen Big Tech’s dominance.

Challenges of Big Tech dominance

• Enormous computational costs: The enormous computational costs of building deep learning models make it nearly impossible for smaller players to compete.
• Deep Learning’s generalised capabilities: Deep learning is now the popular form of AI because it has generalised capabilities.
  • But this is precisely also what makes it computationally expensive.
• Gemini ultra example (2023): As of 2023, Gemini Ultra was the most costly model, costing about $200 million to train.
• Dependence on big tech: To make it affordable, any new entrant would necessarily be beholden to Big Tech companies for compute credits.
• Big Tech’s strategic advantage: These costs also incentivise Big Tech companies to keep advocating for deep learning as the future and pushing out larger and larger models — it locks in their position as the dominant actors and provides the primary revenue stream through which they can recover their costs.

Policy Proposals and Challenges

• Public compute infrastructure: Some recent policy proposals suggest investing in public compute infrastructure or developing a federated model, taking a leaf out of India’s Digital Public Infrastructure model.
• Competitiveness of infrastructure: However, it is not enough just to provide alternate infrastructure.
  • This infrastructure also has to be competitive with the Big Tech offering.
• Big Tech’s developer tools: Big Tech companies offer a wide range of developer tools which make workflows easier and more efficient, and these tools are optimised for their cloud infrastructure.
• Cloud infrastructure and algorithmic models: Along with access to cloud infrastructure, they give companies access to the latest algorithmic models, making tasks such as image or video analysis far easier, along with tools to simplify data preparation and labelling.
• End-to-End service offering: Big Tech’s end-to-end service offering makes development cheaper and easier and drives up the costs of switching to other providers.

Data Monopoly and Its Impacts

• Big Tech’s data monopoly: Big Tech’s data monopoly is even harder to contend with. These companies access a continuous data stream across various domains, social interactions and geographies.
• Sophisticated data intelligence: This “data intelligence” is likely to be more sophisticated than what other players can achieve, giving them a substantial competitive edge.
• Smaller companies’ endgame: Unsurprisingly, many smaller AI companies find their end game is to sell to Big Tech, further entrenching the cycle of dominance.
• Limitations of public data initiatives: While public data initiatives aim to democratise data access and create a more equitable playing field, they often fall short.
• Commercial capture of open data: Open data initiatives are prone to commercial capture, where the better-resourced actors — here, Big Tech with its advanced computational infrastructure and data intelligence — are positioned to best leverage these open data architectures.

Shift in AI Research and Dominance of Big Tech

• Commercial firms dominate AI: The shift toward deep learning as the most popular form of AI has also meant that commercial firms, particularly Big Tech, now dominate AI, and academia has a diminishing role.
• Big Tech’s influence on AI research: Industry players now have more academic publications and citations and are shaping the direction of AI research.

Prioritising a theory of change

• Rejection of the big tech model: We need a radically different approach to AI development that does not aim to compete with or replicate the Big Tech model but changes the rules of the game altogether.
• Problems with the current model: As long as we are locked into a ‘big-data’ and ‘larger is better’ imagination of AI, we will only keep chipping away at an exploitative model of commercial surveillance and even a wasting of precious public resources.

Theory of Change as the Foundation for AI Development

• Theory of Change: A model of AI development whose starting point is a theory of change, i.e., understanding the causal mechanisms through which various factors link together and developing hypotheses about how potential interventions may contribute to change.
• Role of Domain Expertise and Lived Experience: In this model, domain expertise and lived experience guide AI development rather than statistical patterns in Big Data alone.
• Purpose-Driven and Smaller Models: This knowledge and experience are harnessed to develop theories of change and build purpose-driven and smaller models that reflect frameworks for progressive change.
• Targeted Data Collection: Data collection is then targeted and curated to test and further refine the theory of change.

Benefits of “Small AI”

• Democratic and Effective AI: By championing “small AI”, firmly anchored in a theory of change, we can carve out a space for AI development that is inherently more democratic and effective.

Historical Precedents for Theory-Driven Models

• Advancements in Other Fields: Historically, significant advancements in medicine, aviation, or weather forecasting typically relied on theory-driven models, where hypothesis testing and scientific rigour in fields such as biology, physics, and chemistry were prioritised over sheer volumes of data.
• Forgetting Past Lessons: In our obsession with ‘bigger is better’ we seem to have forgotten this entirely.

Conclusion: Another missed opportunity

We need to change course urgently, and we cannot do that as long as we keep viewing Big Data and deep learning as the holy grail. On this current path, we only increase our dependence on Big Tech. The recently signed Global Development Compact is a missed opportunity to re-think the current paradigm. While it makes all the right noises about democratising AI, it ultimately falls back into the same trap of assuming that if countries build large enough data sets and are given access to computational power, we will magically be able to achieve the Sustainable Development Goals and address Big Tech monopolies.

---

Editorial 2: Urgent deadline

Context

Poverty and climate change must be tackled urgently by G-20.

Introduction

The recent G-20 summit in Rio de Janeiro aimed to address global hunger, poverty, and climate justice. Despite Brazilian President Lula Da Silva’s call for taxing the wealthiest to combat poverty, the summit’s focus was diluted by global conflicts. The Global South continues to push for stronger action and governance.

Tackling Global Hunger and Poverty

• G-20 Summit goals: Tackling global hunger and poverty and promoting climate justice were declared goals for the recent G-20 summit in Rio de Janeiro.
• Brazilian President’s call: Brazilian President Lula Da Silva called poverty a “scourge that shames humanity,” asking the gathered nations to implement policies such as taxing the ‘super-rich’, using a 2% wealth tax on the world’s wealthiest to generate more than $200 billion in revenue.
• G-20 Declaration: But the G-20 declaration fell short of that.

Global South’s Concerns

• Prime Minister Modi’s Statement: Prime Minister Narendra Modi too underlined that the problems of the world are felt most acutely by the ‘Global South’, and,
  • therefore, that the reins of global administration must belong to those that represent the larger majority in the world.

G-20 Hosts and Focus

• G-20 hosting by Brazil: The G-20 hosted by Brazil, was by the third host country of the Global South, after Indonesia in 2022 and India in 2023.
• Next Summit location: The next G-20 is to be in South Africa.
• Expected focus: The Brazil summit was expected to focus on solutions for the poorer, emerging economies.

External Factors Impacting the Summit

• Diluted focus: However, its timing diluted the cause and diffused the focus, given the other issues the world confronts.
• Context of clobal conflicts: This was the first G-20 summit since the October 7 attacks on Israel and its reprisals on Gaza and Lebanon.
• Russia-Ukraine Conflict: Russia’s invasion of Ukraine had also made forging consensus at Bali and New Delhi already quite difficult.
• G-20 Declaration on conflicts: With deepening polarised narratives over both conflicts, the G-20 declaration was watered down,
  • expressing only “deep concern” over the humanitarian situation in Gaza, and
  • dropping all reference to Russia while highlighting the “suffering… with regard to global food and energy security.”
  • It was devoid of specifics on ending the conflicts.

Climate Justice and Financing

• COP29 timing: The G-20 was also timed closely with the COP29 in Azerbaijan — Brazil will be in 2025 host —
  • indicating that issues of climate financing and climate justice,
  • which have been raised by the developing world, would find place in the G-20 declarations, and then feed into the COP process.
• U.S. election Influence: However, the summit followed just after the U.S. presidential election results, casting its shadow.
• Trump’s impact: Given his moves during his first tenure, Donald Trump will not set much store by the aspirations of the Global South.
  • Nor is he likely to expend the kind of resources expected from the U.S. towards tackling global warming or in curtailing the exploitation of fossil fuels.
  • His cabinet has climate deniers and his own campaign slogan was “Drill, baby, drill.”

Conclusion: Future G-20 and Global South Leadership

Given the portents, the Global South, and the quartet of Indonesia-India-Brazil-South Africa, will have to ensure that the next G-20 is able to concretise the concerns of the developing world, and set out a path for the future on poverty and hunger, climate change and global governance. In 2026, as the G-20 will return to the U.S., the deadline is more urgent.