Experts discuss impact on jobs, skills requirements and the need for urgent regulation
Artificial Intelligence (AI) is intrinsic to our day-to-day lives today. It helps access our phones or a place through face recognition. When Facebook suggests friends or Netflix recommends what to watch next based on our viewing activity, that is AI in play.
In various sectors of governance, it is used in sectors like telecommunication, banking, finance, healthcare, education, agriculture and defence among so many other areas. AI is being used to conduct elections though electronic voting system, biometric identification, facial recognition, criminal investigation, crowd and traffic management etc. to improve service delivery, enhance citizen engagement and optimise resource utilisation.
In simple words, AI can be defined as a machine's ability to perform a task that would have previously required human intelligence.
The launch of generative AI model ChatGPT in November 2022 has become the most notable advancements of AI, setting the internet on fire in 2023. Its addictive nature has ushered in a paradigm shift not only in manner technology is developing but also how it is affecting human lives making humans completely dependent on it for their work.
Even though AI is a massive enabler, it is evolving and being adopted at a very fast rate. At the same tome concerns are being raised on its safety, privacy, and impact on jobs. Especially for India with huge unemployment in unorganised as well as organised sector, how do you prepare the workforce with AI skills?
In the latest episode of #Checks&Balances, Geetanjali Minhas of Governance Now and a select panel of experts delved into such key concerns with economists, academicians and legal minds.
You can watch the episode here: https://youtu.be/RyCUFSw0xtI
Explaining the difference between automation and AI with examples, Swaminathan Ramanathan, researcher and teacher at Uppsala University, said the two are separate processes. The linking of Aadhaar and different benefits is process automation of the highest order bringing efficiency and effectiveness into people's life, whereas AI is a system that has the ability to think for itself and different levels. And the holy grail of AI is a system that thinks like a human being and acts like a human being independently without any intervention.
He explained that with our underground water table decreasing dramatically and 90% -95% water being used for agriculture, ChatGPT has been deployed on scale by farmers at the ground level, where a farmer is actually able to talk to ChatGPT and relate to the technology in his own language and accent and get relevant information. Here, AI technology is not only helping the farmer with precise water requirements with irrigation and nutrigation but also saving a lot of water which would otherwise get wasted.
Rajiv Kumar, economist, former vice chairman, Niti Ayog, and chairman, Pahle India Foundation, said, “Genie is out of the bottle. AI is all pervasive and its impact is as significant as discovery of electricity in our history.” It will replace jobs maybe in manufacturing where combined with 3D printing it will completely transform the away we manufacturer our product etc.
“Jobs which have to go will go. But there are number of other sectors where it could be used to improve productivity and inefficiency and even the employment levels in the country. For economy like India, which is at its very early stage of economic development and has to become middle-income and high-income economy with a large population to educate and skilling, AI can help absorb our huge requirements, especially to meet the challenges that the country is facing. For advanced economies that have reached a level of kind of saturation in their consumption levels, AI poses a real threat. AI will be absorbed to improve our productivity, efficiency, and most importantly to address the challenge of growing while remaining greener. We need to be we need to be very conscious of this. We need to be careful with this regard, be complacent at all about AI, but in the same time, we have to design our policies towards AI.”
Swaminathan added that jobs with clearly set out processes, high degree of repetition that are currently being done by humans will be replaced by AI as will those jobs that require augmentation.
“I see AI as actually going ahead and working with the human. I think it is more and enabling force in other than disabling force. It will be a system which the human beings will have to start learning how to work with. Think of AI as a co-worker.”
Giving his view on the impact of AI on Jobs and skills required, Rajiv Kumar emphasised on constant upgradation of skills and said. “Period where you got into one career and stayed with it forever is over. In US, people change careers about 3.6 times in their lifetime; the same is going to increase much more. It requires ability for the person to keep learning and be open to learning and be very good at learning. You don't know when the skill requirement will change and how it will change. There is no way that you can anticipate or pre-empt skills required 10 years from now and skilling may always be lagging behind and as they may be teaching you the skills for yesterday.”
He underlined that it is time to give more emphasis on apprenticeship as that will give you the jobs for tomorrow rather than skilling. Kumar also said rote learning must go as it only produces clerks or the lifetime.
Here Swaminathan said that in case of jobs that are going to vanish essentially their core competence has to be maintained while working with AI as co-partners. At the second level AI has to be looked at from a pedagogical perspective and introduced at the school itself.
At the third level, boundary conditions for AI will have to be set up ensuring ethical and moral systems, rules and regulations which will have to be done by humans and not another AI. Here you will need to ensure that the system is intelligent enough to benefit humans and not the other way around, where the system starts controlling humans.
Next, he said, governments also need to take into account and prepare for a scenario where it will be challenging for certain segments of the population to upskill themselves beyond the point. “There has to be a certain social support infrastructure that the government needs to think of, to ensure the people are supported in some way or in the form of a basic income.”
Additionally, the critical issue of cybersecurity of AI itself is bound to throw up new jobs.
Expressing serious concerns about the possibility of human extinction and severe socioeconomic challenges posed by AI, global tech leaders such as Sam Altman, Geoffrey Hinton, Steve Woznaik and Elon Musk have called for a six-month halt on development of AI systems more advanced than Chat GPT urging policymakers to equate at par, risks posed by AI with pandemics and nuclear war.
Renowned Supreme Court of India advocate Pavan Duggal said the biggest problem ChatGPT faces today is the issue of hallucination. Despite Open AI, the company that created the technology, stating in its terms and conditions that this technology in progress and bound to hallucinate and give you wrong answers, please do not rely on it; people are getting addicted to technology and not doing their checks and balances.
Importantly, the output that is coming up from ChatGPT is infringing intellectual property rights and one could potentially see themselves facing some legal trouble and issue.
He added that now ChatGPT is being extensively being misused by prompt command engineering to force it to give answers which are in violation of law. It is being used by all cybercriminals to come up with new innovative mechanisms for committing cybercrimes, or coming up with new mechanism to trouble people, generate more money from potential victims.
Making pertinent points, Duggal said while there is an urgent need to regulate AI fundamentally, how do we recognise AI? Is it a company? Is it a principal? Agent? Can I give it the same kind of rights and obligations that I will give to a legal personality… maybe a company? Maybe a society? What will be the ramifications?
Assuming that you give legality to AI, how do you go ahead and imbibe ethical value to AI because there is bias in ethical standards and compliance where GPT is concerned. We are now increasingly seeing that because they have been actually trained on data sets from the western world, there is a bias in favour of the Western world and against the coloured skin people.
“The fundamental fear of all lawmakers is that in case of AI is not regulated, it could supersede human intelligence at a very early stage. One study tells us that the human elements will be superseded by AI by 2062. I believe that is inaccurate. We should be seeing that tipping point by 2035.”
Rajiv Kumar added that regulation with a light touch and focus on developmental use of AI is important to encourage innovators to use AI for productive and socially useful purposes, rather than let the market take the lead and commercial interest take over.
He said regulation must be framed with very open form of consultation between the different stakeholders and not created in some rooms or by some officer etc and the government.
“You need a conversation between all the stakeholders to come together and then the outline within the structure of their regulations should come through rather than in a top-down manner. A universal, global platform where everybody needs to come together, especially with the warnings of pioneers saying go no further because we cannot as species allow things that are existential threat to their own species as a whole. For that you would require far more cooperation and collaboration across countries and across stakeholders than what is available today.”
Duggal noted that the quandary before the lawmakers will be how to legally recognise AI as legal entity. Also, as AI is cognitive intelligence of a machine somewhere the issue of liability will have to be considered by the Indian law. What happens if a legal injury or legal harm is caused by relying on AI, and if that happens who's going to be held liable for the same? Will it be the company that is selling an AI product or an algorithm? Or will it be the person who's using the AI algorithm or the person who’s actually coded the AI? “So in that sense India will have to increasingly start making the coders liable because once they are coding, the law expects them to have reasonable duty of exercising due diligence to ensure that AI programs do not ultimately come up in a scenario that lands up prejudicing the customers or the users interests.”
He added that AI as algorithm will also have to be audited to make sure it does not capability of becoming rogue or being misused against human interests. He said it will have to be governed though a dedicated legislation or amendment of IT Act 2000 or even through emerging legislation like Digital India. “With cybercrime already on the dark net as a service covered by AI, I expect that India should also come up with certain provisions to penalise AI enabled cybercrime.”