Regulators must learn from past mistakes and act swiftly to prevent Big Tech from monopolising the AI ecosystem, argue Payal Malik and Nikita Jain.
The rise of generative artificial intelligence (GenAI) has been impressive.
Some numbers that illustrate the speed at which technology has disrupted the status quo are in order: To reach 100 million users, the telephone took 75 years, the mobile phone took 16 years, the web took seven years, Facebook took four years, and Instagram took three years. Meanwhile, ChatGPT took just two months.
These estimates underscore the rapid adoption of these technologies and emphasise the importance of getting competition right from the very beginning.
Disruptive technologies always come with the promise of dislodging entrenched incumbents. To facilitate this, competition agencies and other policy instruments should be quick in their response.
In this context, it is important to ask whether the new entrants have the potential to become competitors of Big Tech companies or if they will be integrated into existing business models and mirror the ecosystems that have been created on mobile platforms.
Ecosystems have emerged as an accepted way to describe the economic activity that develops around Big Tech firms as they leverage critical input -- data being the most important -- to orchestrate the creation and provision of products and services.
Are AI platforms that bring together data sets, cloud, computing power, graphics processing units (GPUs), and applications emerging into a new ecosystem?
The focal point of the AI ecosystem is the foundational model, where few firms have the necessary resources, large datasets, infrastructure, and expertise to develop.
It is not hard to imagine AI having a future similar to digital markets, where incumbents monopolise essential upstream inputs (cloud infrastructure and foundation models) and exercise market power over complementors in the application market of products and services.
However, this need not be the case if the trajectory of this new technology is shaped by prudent and quick interventions to stymie any possible anti-competitive conduct such as tying, self-preferencing, acquisitions of complementors or potential competitors, or any other now-known conduct of Big Tech used to leverage their dominance from one market to another.
As AI is still in its early stages of development, it is difficult to say with certainty whether its foundation will eventually become concentrated, but it is crucial to act now.
Competition authorities in France, The Netherlands, and the UK are investigating monopolistic behaviour in the markets for AI infrastructure.
Both globally and in India, the infrastructure (hardware and cloud platforms) and foundational model layers of the AI stack are dominated by Big Tech.
The focus of Indian AI startups has largely been on the applications layer, with the majority working in the code and data segment.
Big Tech incumbents have an advantage in AI due to their access to specialised hardware equipped with a large number of central processing units (CPUs) and GPUs, creating barriers to entry.
The expenses involved in developing and training these models also tend to favour larger firms.
The lack of high-quality, large-scale, localised training data; native hardware original equipment manufacturers (OEMs); and scarce patient capital for cost-intensive foundational model work is likely to increase dependencies of the GenAI startup ecosystem in India on the high-cost computing resources of global cloud companies.
The next competition challenge of AI systems is the possibility for firms to begin with an open-source approach to attract customers; appropriate their competitively-sensitive data and eventually restrict access to their systems to prevent competition -- a scenario reminiscent of what we saw in the mobile ecosystem with the Android Open-Source Project.
Large technology companies have a demonstrated history of capturing or using open-source projects to their benefit, often to entrench and expand centralised power.
There is a growing concern about the increasing partnerships between Big Tech firms and AI startups, as it may lead to a lack of fair competition in the nascent GenAI market, with these partnerships seemingly serving the same function as previous 'killer acquisitions' or 'reverse killer acquisitions' used to create ecosystems.
Interestingly, these agreements may have dodged existing merger regimes. For instance, Microsoft, which has thrown its weight behind OpenAI, does not own any conventional equity shareholding in the company; instead, it is entitled to receive a share of its profits from a specific subsidiary of OpenAI.
In late January, the US Federal Trade Commission announced that it had issued orders to five companies, requiring them to provide information regarding recent investments and partnerships involving GenAI companies and major cloud service providers.
Competition authorities are familiar with the 'playbook' used by digital companies to expand their dominance.
Regulators must learn from past mistakes and act swiftly to prevent Big Tech from monopolising the AI ecosystem.
We cannot afford to let history repeat itself.
Payal Malik and Nikita Jain are professor (visiting) and fellow (consultant), Indian Council for Research on International Economic Relations. Shiva Kanwar, research associate, and Bhargavee Das, research assistant contributed to the column. The views are personal.
Feature Presentation: Ashish Narsale/Rediff.com
