Celebrate AI with caution

SigTuple a healthcare startup based in India has managed to raise $19 million (Rs130 crore) from more than seven venture capital firms recently. The company uses Artificial Intelligence (AI) to diagnose cancer and other dreaded diseases that have been putting enormous economic burden on the nation. Lately, this venture capital fund infusion into AI-led startups has grown exponentially in Asia.

It’s not just the business ecosystem that AI has penetrated.

Countries around the world having become aware of its potential economic and social benefits have put AI at the core of their policy. In the last two years USA, France, Japan, China and UK announced their national AI policies.

While China estimates a 26% contribution to its GDP from AI related businesses, UK puts it at 10%.
Not wanting to lose the opportunity after the Economic Survey 2017–18 highlighted the value of AI, the Government of India mandated NITI Aayog to develop a national programme on Artificial Intelligence. In June this year, NITI Aayog published a discussion paper with a unique brand name #AIforAll.

NITI Aayog's paper explores AI applications for all development sectors like farming, healthcare, education, infrastructure, mobility and defense. The paper also elaborates challenges related to funding, infrastructure, talent creation and regulatory environment, etc.

Stay alert to AI’s dark side
While we should celebrate India's AI vision, it’s important to stay alert to some universal concerns relating to AI and to those that are specific to India’s complexity and the articulated policy.

Let me elaborate them..
Some of us followed the story where Facebook had to shut down its AI systems after ‘chatbots’ created by the company started speaking in their own language. This language was beyond the comprehension of human beings. It was later understood that these AI-powered ‘chatbots’ actually developed their own language!
Beyond human understanding also means beyond ‘human control.’

A few weeks ago, experts on military artificial intelligence from more than 80 world governments converged at the UN office in Geneva for a week’s discussions on LAW (Lethal Autonomous Weapons). The world’s most powerful nations have already embarked on this deadly arms race involving ‘killer robots' and unmanned drones while the regulatory framework is still nascent.

Though India chaired a (CCW) Conventional Weapons Conference at New York in November 2017, we cannot celebrate yet. We as a nation should continue to play a 'decisive role' in international discussions relating to AI regulation.

Calibrate AI to our ecosystem
The AI policy for the nation has to be seen thru the prism of India’s unique developmental challenges. Potholes or garbage overflow detection using AI will not be useful in India as ‘fixing’ those problems is more to do with municipal capacities and not just detection per se.

In the context of farming, usage of certain AI technologies that are popular in Australia may well be irrelevant in India due to the size of the farming units. This is just an example but what is really important is to factor in the country’s vulnerabilities.

Lately, India has been a victim of FAKE NEWS. Experts are projecting AI as a solution for fake news but paradoxically AI can be exceptionally good at actually creating fake news! Using AI, it’s easy to amplify certain biases in the society. This can relate to gender and communal stereotypes and certain sections of society. It can even have political consequences – potentially restricting expression of particular viewpoints and amplifying others.

Machine learning systems are capable of creating 'Deep Fakes’ – photos and videos that realistically replace one person’s face with another. Create material that never existed in the first place, publicise videos of a public figure saying things he’d never actually said. Scary but true.

The price we will pay for not contextualizing AI policy to India’s diverse, layered and extremely complex ecosystem

Let me draw your attention to another important aspect... The NITI Aayog paper says “India provides a perfect ‘Playground’ for enterprises and institutions globally to develop scalable solutions which can be easily implemented in the rest of the developing economies." So far, as a nation, India has been a great ‘Marketplace’ for many a global product but this ‘play ground’ implies that we may become a ‘marketplace for experimentation.'

The country with its vast diversity and a thriving yet complex eco system will be an alluring 'Test Bed for AI’. We should be wary of the world’s temptation to run ‘clinical trials in AI'as India happens to be a vast terrain of opportunity for them.

This aspect cannot be over emphasized as India, according to experts, has exponential power to adopt disruptive technologies. It is predicted that adoption of cognitive technologies in developing countries like India would be much faster than in developed nations, and the magnitude of change it will bring will be far far larger.

Needless to say anything that gets introduced in India, the adoption rate here will be wayhigher than USA, China, Japan and the rest. Reason enough to tune into the wiser counsel of Stephen Hawking, Jack Maa and Elon Musk who spoke on dangers of AI.

As an extremely agile economy, poised for exponential growth we can celebrate AI but with caution.

By:Karuna Gopal

(Writer is President, Foundation for Futuristic Cities)