13 minute read
Manufacturing in India
from Regulatory Sandboxes for Artificial Intelligence: Techno-Legal Approaches for India, ISAIL-TR-002
company Maruti Suzuki has been using artificial intelligence along with internet of things and robotic process automation to automate and increase efficiency in automobile manufacturing processes.65 Tata Group has partnered with the company AEye to develop an autonomous taxi in India called “RoboTaxi”.66 Hyundai India has launched artificially intelligent DAL-e (Drive you, Assist you, Link with you- experience) that uses artificial intelligence and natural language processing to automate vehicle crash warnings, provide emergency assistance and smoothen driving experience.67 Mahindra & Mahindra has integrated Amazon’s Alexa in its vehicles which use artificial intelligence for navigation, vehicle control, playing music, audiobooks, control doors, windows, check cabin temperature, check fuel level, check tyre pressure, check traffic and find parking.68 The two- wheeler company Revolt has launched its artificially intelligent bike which uses artificial intelligence technologies for providing information on vehicle range, real time battery checking, geofencing and to run real time diagnostics.69
Artificial Intelligence in the Automation of Manufacturing in India
Advertisement
Automation of manufacturing is one of the key areas where the deployment of artificial intelligence has been a success and India is no exception to such uses. Companies such as TVS Motor Company Ltd., Asian Paints Ltd., Blue Star Ltd. and JK Tyre & Industries Ltd. have been using artificial intelligence based solutions for undertaking manufacturing.70 Blue Star Ltd. has collaborated with Universal Robots to deploy artificially
65 Ibid. 66 Ibid. 67 Ibid. 68 Ibid. 69 Ibid. 70 BusinessWire, ‘Artificial Intelligence in the Indian Manufacturing Industry, 2020: Trends, Insights, Use Cases, Case Studies’ (BusinessWire, 10 April 2022) <https://www.businesswire.com/news/home/20200428005405/en/ Artificial-Intelligence-in-the-Indian-Manufacturing-Industry-2020Trends-Insights-Use-Cases-Case-Studies--ResearchAndMarkets.com> accessed 11 April 2022.
Regulatory Sandboxes for Artificial Intelligence: Techno-Legal Approaches for India, ISAIL-TR-002
intelligent robots for expansion of copper tube expansion which has significantly resulted in increase in safety at their manufacturing unit.71 A multinational company called GreyOrange has established its research and development centre in India in order to design, manufacture and install artificially intelligent robotic systems at warehouses.72 The company Panasonic has built Technopark in India which harnesses the power of artificial intelligence to undertake production and testing of washing machines and air conditioners.73
The Centre for Product Design and Manufacturing in collaboration with the Indian Institute of Sciences and Boeing are developing India’s first fully automated and artificially intelligent factory which will not only be able to gather and furnish real time data but also undertake planning, monitoring, detection and improvement of manufacturing processes.74 Similarly, Bharat Heavy Electricals Ltd. has invested in initiatives and research to implement and integrate artificial intelligence for manufacturing process monitoring, manufacturing process controls and for maintenance related activities.75 The company Bosch uses artificial intelligence in its manufacturing unit to reduce the time taken to calibrate pumps for tractors. 76 Several initiatives have also been taken by the Indian Government to ensure that innovators get a chance to undertake artificial intelligence development activities related to the manufacturing as well as other sectors in India. These include the Digital India initiative, establishment of Centre of Excellence for Data Science & Artificial Intelligence, Digital Agriculture Initiative, Policy on Cloud Hub and Artificial Intelligence and Task Force on Artificial Intelligence.77
71 Ibid. 72 Geethanjali Jujjavarapu et al., ‘AI and the Manufacturing and Services Industry in India’ (The Centre for Internet & Society, 2022) <https://cis-india.org/internetgovernance/files/AIManufacturingandServices_Report_02.pdf> accessed 11 April 2022. 73 Ibid. 74 Ibid. 75 Ibid. 76 Ibid. 77 Ibid.
[ 29 ]
Artificial Intelligence in Surveillance, Internal Security and Defence on Land, Sea and Air
Artificial intelligence is not only being used for external security related purposes such as by the military, navy and the air force, but also for the preservation of internal security and surveillance by the local police in India. The Delhi Police has used artificial intelligence in the past to apprehend and arrest suspects of the Delhi Riots.78 The Kolkata Police has invested in installing CCTV cameras that will use artificial intelligence to spot illegal parking and bikers without helmets.79 Telangana is building an Integrated Police Command Centre which will use artificial intelligence to undertake surveillance in the State.80 The Kerala Police has introduced an initiative to train its police personnel in the State to use artificial intelligence based analytics tools in order to reduce the State’s police workload.81 The Indian Institute of Technology Kanpur has introduced the Artificial Intelligence and Entrepreneurship Program which is developing an artificially intelligent search engine to help police departments in predictive policing and crime mapping.82 The Bengaluru Police have invested in an initiative which will help in the introduction of mannequins enabled by an artificial intelligence software that will notify the police about criminal activity and about traffic violations and the like.83 A startup
78 Gaurav Jain and Raghav Chopra, ‘AI Tech is Increasingly Being Used by Police Worldwide: Here’s Why India Needs to Regulate it’ (Scroll, 01 April 2021) <https://scroll.in/article/989094/ai-tech-isincreasingly-being-used-by-police-worldwide-heres-why-india-needsto-regulate-it> accessed 12 April 2022. 79 Poulomi Chatterjee, ‘How Helpful is the Growing AI- Based Policing in India’ (Analytics India Magazine, 22 March 2022) <https://analyticsindiamag.com/how-helpful-is-the-growing-aibased-policing-in-india/> accessed 12 April 2022. 80 Ibid. 81 Ibid. 82 Ibid. 83 IANS, ‘Bengaluru Police to Introduce Artificial IntelligencePowered Mannequins to Fight Crimes’ (The New Indian Express, 22
Regulatory Sandboxes for Artificial Intelligence: Techno-Legal Approaches for India, ISAIL-TR-002
company called Staqu is utilising artificial intelligence to assist law enforcement agencies in India to undertake artificially intelligent predictive policing.84 Staqu’s technology has been utilised by Police Departments of several Indian States such as Uttar Pradesh, Rajasthan and Punjab.85
The Indian Navy has taken an initiative to establish a Centre of Excellence and a lab on artificial intelligence and big data analytics and also has around 30 artificial intelligence related projects currently being developed.86 Similarly, even the Indian Air Force has introduced several artificial intelligence based projects that will help it in improving communication networks, data collection and transmission, automate aerial vehicles and airplanes and undertake predictive maintenance.87 The Indian Army in collaboration with the National Security Council Secretariat have jointly established the Quantum and Artificial Intelligence Lab at the Military College of Telecommunication Engineering in Madhya Pradesh to conduct training and promote research in the field of artificial intelligence and quantum computing and its applications in military and defence
February 2020) <https://www.newindianexpress.com/cities/bengaluru/2020/feb/22 /bengaluru-police-to-introduce-artificial-intelligence-poweredmannequins-to-fight-crimes-2107068.html> accessed 12 April 2022. 84 Abhishek Baxi, ‘Law Enforcement Agencies in India are Using Artificial Intelligence to Nab Criminals- Here’s How’ (Forbes, 28 September 2018) <https://www.forbes.com/sites/baxiabhishek/2018/09/28/lawenforcement-agencies-in-india-are-using-artificial-intelligence-to-nabcriminals-heres-how/?sh=59d3d785241d> accessed 12 April 2022. 85 Ibid. 86 The New Indian Express, ‘Indian Navy Ropes in New- Age Tech With 30 Artificial Intelligence Projects in the Works’ (The New Indian Express, 28 January 2022) <https://www.newindianexpress.com/nation/2022/jan/28/indiannavy-ropes-in-new-age-tech-with30-artificial-intelligence-projects-inthe-works-2412338.html> accessed 12 April 2022. 87 Milind Kulshreshtha, ‘AI for IAF to Enhance Air Combat Effectiveness’ (Salute, 15 April 2021) <https://salute.co.in/ai-for-iafto-enhance-air-combat-effectiveness/> accessed 12 April 2022.
[ 31 ]
activities.88 The Defense Research and Development Organisation through its Centre for Artificial Intelligence and Robotics has undertaken several projects to integrate robotics with artificial intelligence and has even launched artificially intelligent robots such as RoboSen.89 Therefore, it can be seen that the potential of usage of artificial intelligence is multifarious in surveillance, internal security and defence of land, sea and air in the Indian context.
88 Press Trust of India, ‘Army Sets Up Quantum Computing Lab, AI Centre at Engineering Institute in Mhow’ (The Indian Express, 31 December 2021) <https://indianexpress.com/article/education/armysets-up-quantum-computing-lab-ai-centre-at-engineering-institute-inmhow-7697802/> accessed 12 April 2022. 89 Satavisa Pati, ‘Use of Artificial Intelligence by Indian Army in the Borders in 2021’ (Analytics Insight, 20 October 2021) <https://www.analyticsinsight.net/use-of-artificial-intelligence-byindian-army-in-the-borders-in-2021/> accessed 12 April 2022.
Regulatory Sandboxes for Artificial Intelligence: Techno-Legal Approaches for India, ISAIL-TR-002
3 Recommendations on Creating a Legal
Technologies in India
Considering the fact that the sectors which have shown immense potential in artificial adoption in India have been discussed and the problems existing in current regulatory sandbox frameworks have also been discussed, it is necessary that certain viable recommendations are provided that attempt to create a legal structure for regulatory sandboxes for artificial intelligence technologies in India.
So far, sectoral regulators have introduced regulatory sandboxes that are restricted to the respective sectors which operate completely within the limits of the powers conferred to such sectoral regulators. Thus, in order to address the dilemma directly, the sector- wise approach will have to change to the technology- wise approach since the sector approach is much narrower than the technology- wise approach. This is because the idea and purpose of a regulatory sandbox is to promote innovation of a new technology rather than to promote sectoral development and by keeping a sector- wise approach, the core purpose of a regulatory sandbox is defeated. Emphasising on this, the shift of approach from sectoral to technological would mean that any legal framework proposed for regulatory sandboxes cannot be formulated through delegated legislation, as is being done currently and therefore, principal legislation will have to be framed. A principal legislation on regulatory sandboxes would harmonise various quintessentials of a regulatory sandbox, something which is not seen currently in the Indian scenario since sectoral regulators have continued to implement sandboxes as per the limitations and possibilities conferred to them through their respective enabling legislations.
Moving forward, the proposed principal legislation would have to stipulate several provisions to achieve harmonisation amongst the necessities of implementing a regulatory sandbox. A
[ 33 ]
provision for definitions will have to mandatorily be kept in order to avoid any ambiguities. This should, ideally be followed by the establishment of a statutory Innovation Office at the Centre and State Innovation Offices along with statutory authorities working in such Innovation Offices, that will assist in the designing, planning, implementation and solving field related problems during the operation of a regulatory sandbox. The qualifications, duties, powers and functions of the statutory Innovation Offices at the Centre and State levels and statutory authorities working in such Innovation Office will have to clearly be outlined.
Considering the fact that a regulatory sandbox will never be limited to one form of technology, especially in the case of artificial intelligence since the utilisation of artificial intelligence is often seen with a combination of internet of things, big data analytics, cloud computing, blockchain and robotics, it is necessary that the principal legislation provides few mandatory provisions and confers liberty for specific regulations, rules, guidelines and frameworks to be made for particular forms of technology as and when regulatory sandboxes for such technological combinations are being implemented. This can be achieved by delegated legislation depending on the objective of the regulatory sandbox and the class of innovators being permitted to participate in the regulatory sandbox. Although eligibility and entry requirements can be prescribed by delegated legislation on the basis of the form of technology being proposed to be tested in the regulatory sandbox, it is necessary that the principal legislation itself provides standard provisions for duration of a regulatory sandbox. It is recommended that all regulatory sandboxes should be implemented and operated in cohorts for a standard period of 12 months (extendable for another period of 6 months only in cases of unforeseen exigencies) and the statutory Innovation Offices keep a check and report the progress of each and every cohort in the 12 months period. Standard provisions for exit from the regulatory sandbox will also have to be provided under this principal legislation. Needless to say, since several procedural portions of the regulatory sandbox will be governed by delegated legislation, the power to make rules, regulations, issue notifications and circulars will have to be conferred by the proposed principal legislation.
Regulatory Sandboxes for Artificial Intelligence: Techno-Legal Approaches for India, ISAIL-TR-002
Although having a minimum insurance requirement will not deter participation in the regulatory sandboxes, at the same time, it is pivotal to have provisions pertaining to incentives in the proposed principal legislation on regulatory sandboxes. The incentives could involve several aspects such as the relaxation of licensing requirements in case suitable outcomes are achieved in the regulatory sandbox, concessions in the payment of electricity duties (a commodity required in insurmountable quantities while testing new technology) and a concessional rate of Goods and Services Tax for supplies made in relation to the testing of the new technology and permitted only during the period of participation of the innovator company during in the regulatory sandbox. In order to ensure that the incentives are utilised only towards the intended purpose, it is necessary that a Concession Agreement is entered between the statutory Innovation Office and the innovator company testing their product in the regulatory sandbox.
Another indispensable set of provisions which the proposed principal legislation will have to deal with, is consumer protection and dispute resolution. In terms of consumer protection, apart from the minimum insurance requirement, a provision mandatorily requiring the innovator companies to take consent from consumers participating in the sandbox framework will have to be provided for along with reporting requirements to be complied by such innovator companies and supervised by the statutory Innovation Offices. Obtaining consent as prescribed by the proposed principal legislation will also help in ensuring that the liabilities of the innovator companies is not unlimited so as to deter them in testing their products effectively. However, there may arise several situations where the damage faced or the transaction entered into between the innovator company and the consumer participating in the sandbox framework are not as simple and therefore, it is necessary for the provisions of the proposed principal legislation to provide for a robust dispute resolution mechanism. It is recommended that a hybrid alternative dispute resolution method of mediation followed by arbitration (commonly referred to as “Med- Arb”) can be used for dispute resolution since the classes of disputes in a regulatory sandbox do not necessarily arise only between the
[ 35 ]
innovator company and the consumer facing damage but also arise between two or more innovator companies participating in the regulatory sandbox framework and between the innovator company and the regulatory sandbox implementing authority. For this purpose, a panel of mediators and arbitrators can be formed as per rules and regulations prescribed in the delegated legislation and necessarily, experts in the fields of artificial intelligence and technology law can be appointed in the panel. This will not only help in achieving outcomes of disputes sooner but also help in ensuring that disputes are being handled by experts in the field of technology and artificial intelligence.
Lastly, considering the fact that the proposed statutory Innovation Offices will have to work in collaboration with the various Ministries of the Central Government, State Government and also with sectoral regulators, an enabling provision allowing Innovation Offices to actively undertake inter- departmental co-operation can be stipulated as well.
In spirit, the recommendations provided above attempt mainly at harmonisation but also aim towards a robust legal framework for enabling technological innovation for taking place. It follows that if statute- governed regulatory sandboxes are implemented in India, the immense unexplored potential of artificial intelligence technologies which only seem to have done well in few sectors can also progress and be applied in other sectors as well leading towards a more quicker yet guided technological innovation in the country.
Regulatory Sandboxes for Artificial Intelligence: Techno-Legal Approaches for India, ISAIL-TR-002
4 Epilogue
The discussions made in the present technical report are clearly indicative of certain ground realities such as firstly; there are inconsistencies between the various regulatory sandbox frameworks in India because of the inherent limitations of sectoral regulators, secondly; there is immense potential for the legal framework of regulatory sandboxes to evolve in India, thirdly; various sectors in India have shown impressive adoption of artificial intelligence technologies but a significant portion remains untapped, and fourthly; a strong legal framework for regulatory sandboxes can help in fostering innovation and launching new artificial intelligence based products in the Indian market and even possibly increase the amount of sectors in which such technologies are used.