Economic Times, June 01, 2020
By Amar Patnaik and Udai S Mehta
Artificial Intelligence (AI), believed to be the most important innovation after electricity, is an integral part of the Fourth Industrial Revolution (4IR). While the global AI market rose by 154% in 2019, it is predicted that 80% of emerging technologies will be based on AI by 2021.
Benefits of AI adoption
AI is closely linked with consumer welfare. Consumers are greatly benefited when AI drives innovation to provide novel interactive applications using facial and voice recognition, speech to text (and vice versa) conversions, virtual assistants, etc. to improve the user experience of availing products/services through price/product comparison tools and tailored products/services. It also maximizes efficiency by reducing information asymmetry between the buyer and the seller and optimizes price discovery, thus, leading to wider consumer choice, and better availability/access to products/services. In addition, it enhances convenience by enabling smart appliances/machines, such as robotic lawnmowers and vacuum cleaners, smart plan watering systems, and other smart home appliances.
It is also known to be having the potential to drive effective delivery of public services, such as those pertaining to healthcare and education, among others. For instance, AI is being extensively used in finding public health solutions for Covid-19, including drugs and vaccines, apart from its use in tracking and tracing.
Problems in the adoption of AI
However, consumers also bear the brunt of its misuse. Firstly, AI systems rely on processing big data sets, which contain a mix of personal data, sensitive personal data, community data, and non-personal data. Many times, such data is prone to breaches and privacy violations. Since the Personal Data Protection Bill 2019 is yet to be passed (despite having its many demerits), there exists a void in regulations governing non-personal data and community data.
Secondly, AI systems are only as good as the data they process. Biased data sets on grounds of race, gender, income, etc. are known to lead to biased AI decision- making leading to discrimination and prejudice towards a certain class of consumers.
Thirdly, AI has the potential to enable businesses to ascertain consumers’ willingness to pay for a given product/service. This may lead to price discrimination, through dynamic pricing, thereby adversely impacting consumer welfare. Also, businesses, empowered with AI, can assess consumer preferences and behavior to orchestrate personalized advertisements, offers, and inducements, which although sometimes beneficial for consumers, can also lead to unduly influencing the choices and reactions of the consumers. This may lead to consumers losing absolute control over their purchase decisions.
Lastly, there are also risks of market collusion (tacitly) between close competitors, leading to loss of choice and higher prices for consumers by the violation of competition laws.
These risks arise because while AI makes predictions for consumers based on their data, but it functions within a ‘blackbox’. This ‘blackbox’ creates information asymmetry for consumers in understanding the processes and design of AI mechanisms through which these technologies come to a decision. For example, a consumer might not be able to trust an AI-driven diagnostic tool as it lacks the assurance factor compared to when the same diagnosis is performed by a medical practitioner.
Taken together, such risks are likely to dent consumer trust on AI-driven products/services. The individual need not be afraid of using AI and in fact, need to have more confidence in AI to ease his living. For this, consumers must be kept at the center of any policy or regulation on AI, including mandatory consultation with consumers in policy/regulation-making processes to win their trust in the emerging ecosystem.
Regulation of AI to build consumer trust
Hence, there is a need to regulate the adoption and use of AI. But the regulations need to match the pace of the dynamically evolving AI-driven digital technology, particularly since the adoption of AI in India is in its infancy. Regulations must be evidence-based and ensure that costs imposed by proposed regulations do not outweigh the intended benefits. There should be optimal regulations, wherein innovation is encouraged, while also enhancing consumer welfare through protection from its allied risks. Thus, light-touch regulations with elements of self-regulation have to be encouraged.
India requires immediately a data protection law that can manage the potential negative externalities arising from the use of AI and ensure individual privacy. A robust competition policy would also be helpful in minimizing risks. There is already a consumer oversight mechanism prescribed under the recently passed new Consumer Protection Act to provide adequate consumer redress. This oversight mechanism as also the data protection regime should have the structural and functional wherewithal to test any AI technology for trust, fairness, accountability, and transparency. Further, it should be actively educating the consumers about AI processes so that information asymmetry between consumers and the seller is reduced.
On top of all these regulations imposed on businesses, businesses will also need to ensure ethical use of AI and be transparent with their business practices. Else they would not only face the sword of hard regulation in the form of full government control but may also lose consumers’ trust. Ethical use of AI can also be promoted through creating impact assessment tools that can measure the effect of AI through a trust index as done in the UK. This will help build more ethical and human-centered AI systems.
(Dr. Amar Patnaik is a Member of Parliament, Rajya Sabha from BJD, Odisha, a former CAG bureaucrat, and he is also a member of JPC on Personal Data Protection Bill 2019. Patnaik has a Masters in Public Management from the Lee Kuan Yew School of Public Policy, Singapore and the Kennedy School of Government, Harvard University, and an academic with a Ph.D. in management and Udai S Mehta is the Deputy Executive Director, CUTS International)
This news item can also be viewed at: