Businesses across industries are increasingly deploying artificial intelligence (AI) to analyze user preferences and personalize user experiences, boost productivity, and fight fraud.
Financial services are using AI to hedge financial risks, while the healthcare industry is leveraging AI to improve diagnosis and treatments as well as to reduce the time for development of vaccines. Marketplaces, eCommerce, retail, travel, gaming, entertainment, social media and many other sectors are all using AI in some way or the other.
How much is too much
This growing use of AI has already transformed the way the global economy works and how businesses interact with their consumers. However, in some cases it is also beginning to infringe on people’s privacy. As a result, there is a growing debate about how much AI is too much.
AI suffers from three major issues at present, namely: privacy, bias and discrimination.
That brings us to the moot question whether AI should be regulated?
Also Read: Five reasons why banks may reject your home loan application
Taming the AI juggernaut
I would recommend regulating AI so that the entities using the technology act responsibly and are held accountable. You may counter my view saying that a move such as this may stifle innovation, pace and growth of the market.
While I agree that AI-driven innovations have opened up new opportunities for businesses, I still maintain it is a growing challenge that is resulting in job cuts, and its skewed algorithms are causing racial and gender-based discrimination. To make matters worse, new AI-driven start-ups are sprouting every day, compute power is increasing steadily, and the amount of consumer data being collected – and shared – is growing exponentially.
In this backdrop, AI can be used to profile people like you and me to such a detail which may well become uncomfortable! And this is no exaggeration. You may have already had such an experience when having bought a simple T-shirt from a store, you are inundated with personalized recommendations from all sorts of ‘brands’ on the types of clothes you could buy based on your physique or height. Caution: This example is just a tip of the iceberg!
Why regulate AI?
By AI regulations, I refer to the development of laws and policies that broadly govern the algorithms which will help promote responsible use of AI and make businesses accountable.
While AI can be extraordinarily useful to understand consumer preferences and influence customer behavior through targeted messaging, mandatory regulations on AI can go a long way in preventing technology from infringing human rights. They can help ensure that technology is used for the benefit of end users instead of negatively affecting their lives.
Start with baby steps
The first step the government can take is to create a simple regulatory framework that defines the several capabilities of AI and identifies the ones that are more susceptible to misuse than the others.
Data is the backbone of AI and businesses must learn to access data while ensuring its privacy, integrity and security. Regulations such as GDPR, PSD2, are already in existence and govern how much data businesses can capture and share.
The next recommendation is elimination of black-box approach through mandatory explainability for AI. This will bring in transparency and a clear explanation of why and how a particular result was reached. The details of the steps involved in the analysis and how data was analyzed will help businesses understand the rationale behind every decision made and review the data in each step to even unearth biased results.
It may not be an easy task to regulate AI, as governments seek suggestions from the industry and businesses that would generally make recommendations that are favorable for them. Therefore, to formulate effective regulations, policymakers must try to strike a balance between the scope of the regulation and the vocabulary used.
Benefits of regulating AI outweigh potential losses I am aware that regulating AI may adversely impact business interests. It may slow down technological growth and suppress competition. However, taking a cue from GDPR, that has been successful in the EU, governments can come together to create more AI-focused regulations and have a positive long-term impact.
I firmly believe that the government must engage in meaningful dialogues with other countries on a common international regulation of AI. Having said that, governments must keep their roles limited and not assume absolute powers, lest it hurt the growth of the AI market, in the long term.
(By Gautam Sinha, CEO of LTFLoW. Views are personal)
Disclaimer: The views are not meant to hurt or derail growth of the AI market but solely aim to protect the interest of consumers’ privacy.