A strong data protection law, such as the one drafted by the Srikrishna Committee, will improve data access and quality, benefitting AI research
There are certain crucial periods which determine the course a country is likely to take in the foreseeable future. India has failed to capitalize and build on the industrial and manufacturing revolutions. Now, the seeds of the next major revolution in the form of big data analytics and Artificial Intelligence (AI), are being sown by major economies around the world. India has the capability to partake in the benefits.
AI has the potential to drastically transform a number of sectors such as transport and traffic management, health, education, agriculture, telecommunications, defence and information technology, to name a few. China is reportedly planning to invest $150 billion over the next decade in AI research and development. France is investing $1.8 billion over the next five years and the foremost private companies and researchers leading the AI race are headquartered or based in the US.
To catch up and potentially lead the AI race, it is imperative for India to form a bold AI policy, along the lines of NITI Aayog’s discussion paper, that focuses on data collection, analytics, research and application—since it is data that fuels the development of AI. Fortunately, India has a strategic advantage in potential access to data due to the large digital footprint of its population. It is important that at this juncture, the Justice Srikrishna Committee has come out with the Personal Data Protection Bill, 2018. We believe this will play a key role in unlocking India’s AI potential.
In our opinion, the building blocks of a robust AI ecosystem must have four pillars. First, access to data needs to be enhanced to give firms dealing with AI more data from a variety of sources. However, while it is important to effectuate greater access to data, it is also necessary to ensure adequate protection is provided to personal data which forms a subset of the data pools that big data researchers may have access to. The Bill is a vital step for ensuring robust protection for individuals whose personal data is used in such research.
For example, its provision on research when exempting researchers from certain obligations under the law requires anonymization of personal data as far as possible. Further, a general duty is imposed on researchers to ensure that the individuals whose data they are collecting and processing are not harmed by the research and that such research is not used for the purpose of targeting them. Additionally, when non-anonymized personal data is collected from individuals, researchers would still be required to ensure the security of the personal data and that it is processed in a fair and reasonable manner.
The Bill also makes researchers subject to data protection impact assessments. These are required when processing personal data involves new technologies or the use of sensitive personal data. For all kinds of research which are not exempted by the Data Protection Authority, researchers would still be required to take consent of individuals and inform them of the purposes for which their data could be used. Such individuals should also be allowed to verify the accuracy of such data.
Apart from these measures, if anonymized data does not suit the requirements of the research, measures such as pseudonymization and setting up of ethics committees may be looked at. Alternatively, other forms of differential privacy may be popularized, whereby such data analysis aspires to learn about a particular data set without revealing personal data about a specific individual by studying aggregated data derived out of pools of personal data that they themselves do not have access to. Not only would these measures safeguard the privacy interests of the individuals, they would also encourage them to share their personal data without fear that it may be misused or disclosed without their consent.
Second, the quality of data that entities developing AI have access to has to be improved. This can be achieved through requiring entities to process data in a structured and accurate manner. The Bill does this by making personal data subject to individual access and correction requests.
Third, investments made to support big data research will need to be bolstered by funding from both the public and private sectors, subsidies, tax rebates and so on. The government could identify and fund research projects, or invite domestic and foreign venture capitalists to invest in sectors that it has identified as best placed to reap the benefits. In this regard, the government can set up designated AI research zones in leading universities and research institutes across India.
Finally, a culture of research should be inculcated within students and greater impetus ought to be given to training and employing big data researchers in educational institutions. Until India is able to ensure a steady stream of peer-recognized research, publications and patents, it will be difficult for it to secure a lead in the AI race.
We believe that by encouraging the growth of big data research while also according strong protection to the personal data which may form part of such research, India will be on its way to becoming a responsible and successful global leader in the AI revolution. This belief is in opposition to general industry opinion that data protection laws like the European Union’s General Data Protection Regulation are a hindrance to research and innovation. But a strong data protection law will instead improve data access and quality, benefitting AI research.
Samraat Basu and Gargi Rohi are research fellows at the Vidhi Centre for Legal Policy, and part of the team which assisted the Justice Srikrishna Committee by providing inputs for the committee report and the Personal Data Protection Bill, 2018.
Comments are welcome at firstname.lastname@example.org
Originally Published: https://www.livemint.com/Opinion/l08Yev0nghHLLjJTYmAzHM/Transforming-India-into-an-AI-leader.html