Artificial Intelligence (AI) has gravitated industries across various sectors where each one is trying to leverage it to their advantage. Normally, AI has clout associated with automation, robotics, intelligent decision making, etc. However, as each coin has two sides, AI too has proved to be a boon or bane in recent years.
There have been allegations across the world, and especially in the US, of election rigging by dissuading the mood of electors using AI-based advertising on social media platforms. The Facebook-Cambridge Analytica scandal which was uncovered in 2018 is a testament to the connivance of tech companies to breach the privacy of gullible users and use the power of AI for ulterior and sometimes illegal motives. India being the largest democracy, has the most to lose from this AI threat.
Recent fatal crashes of Boeing 737 Max are another example of the AI-based system going haywire. It is believed that the root cause of these crashes was the failure of an AI-based software component known as Maneuvering Characteristics Augmentation System (MCAS) which was designed to keep the aeroplane afloat under certain extreme circumstances. Boeing trusted this system more than humans, so it gave it higher precedence and during the fateful flights, the pilots just struggled to regain control from faulty MCAS which led to the unfortunate but avoidable disasters.
Stakes for India are very high since it operates one of the largest fleets of modern passenger aircraft that are highly dependent on their AI systems for normal operations. Another AI threat that needs to be tackled is the surge of realistic ‘Deep fake’ photos and videos which are artificially created by software using AI but are indistinguishable from real photos and videos.
With the recent push by the Indian government for autonomous vehicles, drones and AI-enabled weapons, another major challenge has surfaced. Although AI systems can achieve human-level competence in many tasks, however, unlike humans they can easily be fooled. Recent researches have shown that just by making small changes in the environment that are indistinguishable to the human eye, these AI systems can be deceived into making shocking mistakes. Just changing a signboard on a street by a miscreant may make an autonomous vehicle cross speed limit, come to a halt or perform other unexpected actions.
As the list of AI innovations increases, so does its loopholes and threats. We must cautiously embrace AI by making our country AI literate. As a nation, to significantly contribute to AI, huge emphasis on AI research funding is required. To keep up with the pace of AI innovations and research, rapid modernisation of curriculum, project-based learning, and practical research is the need of the hour. A strong workforce focussed on AI research will bring India at the center-stage of the global AI community.
Bennett University (BU) is a state-of-the-art university focussed on AI research and innovation where we aim to make India an AI superpower. Some notable research projects at BU specially focussed on mitigating AI threats include detecting deep fake photos and videos, detecting unusual advertising for ulterior motives, fake news detection, improving AI in autonomous vehicles for Indian conditions, compressing highly accurate supercomputer-based AI models to be used in resource-restricted devices, AI-enabled drones for automated detection of vehicular accidents and other mishaps, etc.
(The author is a research fellow in CSE department at Bennett University)