The laws always lag behind technology; it is time that our laws started catching up
There has been a disturbing trend of late of AI-generated deepfakes circulating the web. This has brought into focus the need to have regulations around AI in India. The government has made regulations addressing deepfakes by AI its priority. Rather than a piecemeal approach, the Government should promulgate comprehensive legislation for addressing the disruption brought by AI.
Such an AI law should be based on the principles of creating and preserving the value of data and content, protecting individuals from harm, preventing market dominance by big players, fixing accountability and liability, and regulating the use of AI by and in the government.
An AI model is largely dependent on access to vast amounts of data to train. Data is not just the new oil but the new gold mine. As such, it is important for a nation like India that generates a gargantuan amount of digital data to create value from this digital asset. The Government can think about building into the AI regulations, a lawful and easy way of creating and sharing datasets as a private-public partnership or private-private partnership.
The AI regulations must also play a role in creating and preserving value for data owners and content creators. In building datasets for AI, little heed is paid to the proprietary rights of the owner whose data is being extracted, many times without their knowledge or permission. This erodes the value of the content created by its authors. An AI regulation needs to draw boundaries around what content may be used to create datasets and how it must be done. Conversely, in generative AI, clarity must be brought as to who is the owner of the content created by AI, under law. The regulations must also answer the question of the patentability of AI. These are important issues to be addressed so that the value of AI can be unlocked fairly for all stakeholders.
AI’s potential to cause harm is well understood. AI can be used to unfairly profile and target individuals, make decisions without human intervention, create and spread misinformation and disinformation, aid in identity theft and generate deepfakes, and cause workforce replacement. While society must keep up with evolving technology, guardrails must be drawn to ensure that dead bodies are not left behind in the rolling juggernaut of AI. India must have a law to prohibit the use of AI for certain purposes and regulate the use of AI for certain purposes. Some accountability measures like subjecting an AI model to safety standards must be prescribed by regulations. Closely linked to the problem of AI’s potential to harm is the question of who should be liable for the output generated or for decisions made by the AI and what should be the threshold for civil and criminal liability.
AI is the new gold rush, and Big Tech already has a head start in the race due to their access to unimaginable amounts of data, their computing capability, and their Large Language Models (LLMs). The Government should think about updating our competition laws so that smaller firms are not left behind in the AI race and can have a piece of the pie too. This will ensure fair competition and prevent market dominance by the usual suspects.
The last bit that the AI regulations must cover is the use of AI by the government. One part will be the government having a policy on using AI to bolster its state capacity, for example, in service delivery, defence, and justice delivery. The other side is to address the use of AI by the government itself, for example, in crime prevention and detection, censorship, etc. If indiscriminate use of AI can be dangerous in private hands, the danger is more potent in the Government’s hands.
The law always lags behind technology. In the case of AI, however, the technology is galloping at an exponential pace. It is time that our laws started catching up. The society must not be in a position where a sentient machine has unchecked power.
(This article is authored by Akshaya Suresh, Partner, JSA)
