Vikas Agarwal, an expert in Artificial Intelligence, Machine Learning, and Cloud Computing, writes a special column for Deccan Mirror about whether a decentralized AI is the need of the hour.
Have you ever wondered how much control a user has over the AI tools?
Question yourself, and after thinking for a while, one can easily understand that all that we see in the AI market is once again like the tools developed during the 2000s, which limit the amount of operability for the users. The scope of AI goes beyond being able to converse with a system regarding information (GPTs) or generating images/videos/audio using user prompts. This is the problem we encounter with AI centralization.
With tech monopolies getting ready to overtake the newly formed AI technologies, it’s again going to be a nightmare for users, as they have to pick only the options that they were offered. The story doesn’t end there; users will be denied and restricted from data that is supposed to be easily available across the AI platforms. So, the philosophy, the limited, diverse data sets, and everything can be controlled by the tech giants once again in the AI ecosystem, to which decentralization offers the solution.
As of 2023, just three companies—Google, Amazon, and Microsoft—account for over 65% of the global cloud computing market, a crucial backbone for AI applications, which directly implies that progress in AI depends on the decisions of the monopoly organizations.
The decentralized AI systems, often referred to as deAI, powered by blockchain and federal learning models, offer an effective alternative to the centralized AIs. By distributing control and ensuring transparency, the major issue of biases can be straightaway fought back, making the AI models more diverse and people-friendly.
Projects like OpenMined and Ocean Protocol are already paving the way for such a future, demonstrating that innovation need not come at the cost of centralization.
The limitation of monopolies’ control over the emerging AIs is not just to be called a people’s movement but needs action from policymakers, too. The robust relegations from the policymakers to restrict the role of large corporations and hinder them from overtaking a major chunk of the contributions will surely pave the path to decentralization. One needs to learn lessons from history. AI machines, which are soon going to transform the human world, should not become just machines of bosses but the friendly tools of humans. Let’s hope decentralization comes close to that.
