Dr Urvashi Aneja (@urvashi_aneja) writes for Chatham House (@ChathamHouse) on the challenges ahead of India's #AIForAll Strategy: despite great promise, the strategy cannot be a panacea for all of India's social and economic needs; it will be constrained, for example, by India's existing data systems – which could be outdated, fragmented, or unrepresentative of the population at large. More than just that, the existence of #DigitalDivides across gender, region, age, class and language means that even existing data, when cleaned, up to date, verified or integrated to the extent possible could contain severe biases in the best of circumstances. But some data challenges may remain even if internet penetration increases, alleviating the #DigitalDivides. For example, one might need new means of #DataCapture and verification. #DataPrivacy will remain a challenge despite a consent-based disclosure regime because *informed* consent needs both IT and basic literacy, and thus practice of ethical AI could conflict with obtaining unbiased or complete data.
But the challenges facing an all-India AI strategy will not stop at Data issues. Implementation of AI in industry could result in labour disruptions, for example if Western economies were to use the potential of AI for intelligent automation in manufacturing, business processes or services, one could have the phenomenon of '#Reshoring' or '#RightShoring' the thematic opposite of 'Offshoring'. This might impact the economic rationale for #MakeInIndia, or indeed, also the BPO/IT industry. Simple BPO work could be automated altogether. At the same time, AI-enabled #Globotics, a possibility in which work could be done remotely (internationally), during a transitional phase (or even longer) present a possible opportunity. Thus the 'AI for All' strategy needs to navigate a tricky future trajectory. via @samirsaran https://t.co/v4rBl05yxE
