Famous English mathematician Alan Turing once said, “A computer would deserve to be called intelligent if it could deceive a human into believing it was human.”
The same thought was encapsulated in the rigorous Turing Test, where a human interrogates an entity. If after completion, the interrogator cannot tell whether the entity is human or machine, the entity is said to be ‘intelligent’. But modern technological research has pushed the boundaries of innovation in intelligent computing, corroborated by the fact that multiple programmes like chatbot Eugene Goostman and OpenAI’s GPT-4 have passed the test.
Today, these intelligent programmes have altered the way we study, research, practice profession, etc. But along with this comes a threat of its potential misuse in disrupting work, lives, society and even democracy.
AI systems have exacerbated societal biases because of the biased data they are fed without any recognition of historical context or social injustice. Computer-aided diagnosis systems have been found to give higher returns for white patients than others. Security datasets used by the police could amplify bias against certain communities. All of these could result in forming parochial narratives.
Further, with the growing use of sophisticated AI systems, various jobs have become redundant, and more will follow suit. This would mostly displace workers from a disadvantaged socio-economic background, exacerbating inequalities. While there is some evidence that AI will create new jobs, the process of skilling for these jobs remains ambiguous.
AI systems have made the phenomenon of deep fakes extremely prevalent, along with growing anonymity. Use of the same by state and non-state actors has serious implications on election outcomes across the globe. In addition to these, issues of copyright infringements, privacy and security also require immediate attention. All such problems require dynamic solutions.
This is where governments and policymakers across the globe must come together to acknowledge and address the threats possessed by AI systems. The goal should be to harness the powers of such systems, while also delicately regulating their use so that they do not impede human rights.
In India, however, there is a lack of dedicated regulations for AI. The Union government has only issued guidelines aimed at responsible development of AI technologies. It continues to follow a majorly non-regulatory approach despite the number of threats. The proliferation of deep fakes and fake news peddled with the use of AI is a serious concern as evident in the recent Lok Sabha elections.
There were 14 AI-generated electoral ads for violence against minorities circulated on social media. This fake propaganda is alarming in a country where 60 percent of households are not adequately digitally literate. But the Centre’s reluctance to act against such instances is not surprising.
Apart from broad claims of making the youth AI-skilled, the Centre should aim to gather information on the vulnerability of various job profiles bound to be impacted by AI. This will help us understand skilling needs better for different workers. According to an estimate, India would need more than one million AI-skilled engineers over the next few years. This seems unlikely unless there is an extensive strategy for the same, bridging academia and industry.
The Centre has also been largely silent on copyright infringement by AI or providing protection for AI-generated content. Moreover, lack of clarity persists on creations that have an overlap of human and AI efforts. Such lacunae only discourages innovation.
Further, it has completely ignored the problem of data bias relating to AI systems. While it continues to promote entrepreneurial action in the AI space, there is absolutely no regulation that encourages transparency in training datasets of AI systems. Lack of diversity in training data will result in perpetuating inequalities by favouring certain demographics and cascade into the quality of output.
The Centre was quick to come up with a fact-check unit to curb dissent on the Internet under the guise of checking fake news, but is still pondering on a comprehensive cohesive AI strategy. The aim for this strategy should be to play between innovation, data protection, risk assessment (along the lines of EU’s AI law) and effective monitoring of such systems.
Tamil Nadu, in recent times, has become a focal point for AI-led initiatives in the country. During Chief Minister M K Stalin’s recent visit to the US, memorandums were undertaken for greater investments in the state by tech giants. To complement this, the state government continues to work on policies for Deep Tech and its ethical concerns.
AI is the future. There is little to discuss there. But building a sustainable AI ecosystem that benefits society, safeguards human rights, promotes innovation and upholds democratic values should be our ultimate goal. Governments should ensure policymakers and regulators have access to the expertise required for regulating, monitoring and enforcing required policies.
Further, policymakers globally must engage in multilateralism to make such regulations interoperable in order to plug all the gaps and ensure AI serves not just as a tool of convenience but also as a tool for equality, justice and opportunities. This will make AI truly ‘intelligent’, for all of us.
Thamizhachi Thangapandian
Member of Lok Sabha from South Chennai
(Views are personal)