Spend on AI by businesses in India is expected to grow at a CAGR of 39 percent and reach $11 billion by 2025.
Can artificial intelligence (AI) be supervised to be responsible and ethical? Can this supervision lead to several high-paying jobs minding and training bots to be responsible and ethical? Is there a need for Responsible and Ethical AI? The answer is a resounding Yes and India has the talent, the market, and the capacity to take a leadership position in the world of Ethical AI.
Recommended ArticlesView All
Budget 2023—ESOP tax incentives drive start-up growth
Jan 31, 2023 IST6 Min(s) Read
Budget 2023: Focused investment in edutech and human capital must become an imperative
Jan 31, 2023 IST2 Min(s) Read
This Microsoft project is enabling the digital presence of low-resource languages
Jan 30, 2023 IST7 Min(s) Read
CEOs, CXOs may see 9.1% salary hike, focus on performance-linked pay: Survey
Jan 30, 2023 IST4 Min(s) Read
AI is everywhere today. Serving you clickbait as you scroll through your Facebook feed and providing you suggestions as you type on Gmail. Spend on AI by businesses in India is expected to grow at a CAGR of 39 percent and reach $11 billion by 2025. Given the expected high growth, it is also important to talk about Responsibility and Ethics in the application of AI.
China is the world’s leader when it comes to research on AI as well as the application of AI for business automation. One such application is in collections on the massive $200 billion debt pile in the country’s peer-to-peer lending industry. A rising wave of defaults resulted in several AI platforms coming up in 2018 to help recover the delinquent loans. These systems scrape the Internet for information on the borrowers and then contact the borrower via phone using a dialog bot which uses the phrasing with the highest likelihood of pressuring the person to pay back the loan. The system also calls friends of the borrower and asks them to relay the urgency of making payments.
Clearly, these methods are working. AI-led debt collection is showing two to four times better results than traditional methods. However, there is a dark side too. Unsupervised algorithms have also been reported in several cases resulting in severe borrower stress including cases of harassment. All this points toward the need for supervision and Ethical AI.
In India and Africa, unregulated lending Apps with roots in China started appearing and grew rapidly during the pandemic. In December alone, India reported two borrowers’ death by suicide in separate incidents allegedly due to harassment by online lenders. Earlier during the year, Kenya also saw a rise in complaints of harassment from online lenders that led SBK, the regulator, to ban a few of them. Even in a more advanced economy like Israel, government regulators had to sound warnings to collectors as threats of suicide by borrowers became more common.
As human well-being—the edifice of the Responsible AI framework is shaken, the believers in the power of Big data and AI hoping to bring a difference in the way credit and finance function are left with tough questions to answer. While a lot of success has been achieved in customer acquisition using advanced data science and complex statistical modelling techniques, the approach has seemed to have failed at most things beyond. This reduction of AI and Big data to nothing more than a sophisticated customer acquisition tool focussed only on the top line and the bottom line is a grave concern.
Among the issues that have emerged from the media reports about the unfortunate incidents around lending Apps is the irresponsible execution of AI —specifically a highly questionable practice where the apps send out harassing messages about non-payment to borrowers' close contacts. While co-obligant and joint liability concepts are not entirely new, their arbitrary application using the power of data and digital technologies is extremely unethical.
The warning signs that have emerged should make the leadership team within all the financial institutions sit up and take serious notice. With enough success achieved on customer acquisition, it's now time to focus energies on creating systems that can serve the financial institutions' best interests and use AI as an effective tool for customer engagement rather than coercive intrusions. AI can indeed be moulded into a great tool for enhanced customer success, although the possibilities of this have not been fully explored.
Ethical AI is a practice of evaluating the ethical quality of the impact of its prediction on human life. Debt collection is a good place to develop this practice because human debt collectors with performance targets typically find it hard to navigate the moral quagmire involved in dealing with financial stress. India, the land of spirituality, is well-positioned to become a leader in the practice of Ethical AI. The day is not far when the practice of Ethical AI will be a key differentiator for AI Platforms.
On the policy side, countries across the globe are also working on Data Protection laws in the model of EU's GDPR, which would give customers the legal Right to Explanation. In the future, an Algorithmic Accountability Reporting framework would also be something that lending institutions may have to work on. While the policy and regulatory efforts will serve as a guide, the responsibility of building Ethical AI systems would lie with the institutions themselves as practicing Data Responsibility and Ethical AI becomes a key differentiator in guiding consumer choices.
—The authors, Raj MKK is Founder & CEO, CreditNirvana and Anup Pai is Founder & CEO, eSamudaay. The views expressed are personal
(Edited by : Ajay Vaishnav)