April 1, 2021
Can technologies based on artificial intelligence (AI) or machine learning (ML) be patented? What are the intellectual property (IP) issues faced by businesses that create products and services based on AI and ML? Experts in a recent webinar organized by the Intellectual Property Facilitation Cell (IPFC) at ICRISAT answered these questions and more.
Mr Rajaraman Srinivasan, Co-founder and Chief Operating Officer, Neoware Technology Solutions, and Ms Akshaya Suresh, partner at VB Legal, shed light on the growing prominence of AI and ML and the role of IP during the webinar ‘Importance of Intellectual Property in Artificial Intelligence and Machine Learning Technologies’.
AI and ML are being utilized to solve complex problems that businesses face. For example, four key areas in agriculture that AI and ML are being used for are automation (autonomous tractors and robots for picking or spraying) to address the shortage of labor; soil moisture detection to improve productivity (crop yield); disease identification to monitor crop health; and precision farming.
“AI and ML are driving important developments in technology and businesses because of the availability of data, scalable computing, skills, investments and awareness,” said Mr Srinivasan.
Ms Suresh, a legal expert on IP management for technology businesses, explained how IP is relevant to AI and ML. She began by suggesting what the objectives of a good IP strategy for AI and ML should be. The four key areas where IP is critical to the use, protection and monetization of AI and ML-based technologies are: training datasets, ML algorithms, software and output.
Training datasets refer to the initial data used to develop an ML model. While using the ML datasets, one must be cognizant of third-party rights of publicly available resources, ownership of the data and regulatory issues. ML algorithms are techniques to solve a problem. In India, protecting an algorithm is challenging because there is a ban on patenting algorithms and computer programs, and copyright laws do not fully protect. The only way to protect algorithms is to treat them as trade secrets.
Before using open-source or licensed software, it is important to read the fine print carefully to understand how they can be used. Ms Suresh went on to explain the different kinds of open-source software licenses available. Both copyright and patents are used to protect software or computer programs. The output of AI or ML varies according to the intended purpose of AI, which means it can be a prediction, recommendation or classification.
“For example, if AI is used in crop sowing, the output would be predictive analytics to determine when and how to sow,” she said. AI outputs cannot be protected using copyright and patents as computers cannot be authors. Contracts are the most feasible option if there are specific aspects of the output that requires protection.
So, what is the way forward for IP in the AI and ML space? The future of sound IP practice in AI and ML space is to have a national innovation ecosystem that includes sound policies, infrastructure, expertise and capital. It also requires companies to adopt an appropriate IP strategy for AI and ML products and services. India is still in a nascent stage in the IP space and must evolve as AI landscape rapidly changes.
Autonomous machines and sensors based on AI and ML are the future of agriculture. Agriculture will increasingly become data-driven, enabled by AI and ML tools. Given the rapid pace of changes in these technologies and the complex nature of the tools, businesses and startups working in this space should create a sound IP strategy to survive and thrive.
The webinar was jointly organized with the Bangalore Bioinnovation Centre on 10 March 2021. A total of 70 participants including entrepreneurs, researchers and innovators participated.
A recording of the webinar can be viewed here: https://bioinnovationcentre.webex.com/recordingservice/sites/bioinnovationcentre/recording/playback/a3bd1fb2457d49f5accffa0cbc3bbb5a
Author
Sridivya Mukpalkar, Consultant – Content Development and Promotion, AIP- ICRISAT
Topics
Species
