The recent buzz regarding ‘ChatGPT’ has got people elated and petrified. Many are still experimenting and forming opinions about its utility. For the uninitiated, ChatGPT, as it is popularly called, belongs to the family of Large Language Models (LLMs). LLMs not only generate human-like text but can also write an essay, summarise content, write poetry, plan travel itinerary, generate computer codes, and even draw a vivid picture.
ChatGPT, an LLM application developed by OpenAI, is capable of contextual learning and answering questions using supervised and reinforcement Artificial Intelligence (AI) algorithms. Early impressions suggest that LLMs can penetrate diverse sectors including Banking, Tourism, IT, Manufacturing, Health, and Education. They can also impact organisational functions such as Marketing, Operations, Accounting, and Communications.
Within two months of its prototype launch on November 30, 2022, ChatGPT had over 10 crore users and nearly 1.3 crore daily visitors, making it one of the fastest-growing consumer applications ever.
While LLMs have received a positive response, many observers have expressed concerns about the negative implications of their usage for society at large. Goldman Sachs has estimated that around 30 crore jobs may be affected globally by the latest wave of Generative AI. Some are pushing to stop feeding more data to such AI systems to avoid future consequences. Some countries have even blocked ChatGPT for data privacy concerns.
The global Generative AI market was estimated to be worth $1.2 billion in 2022. A CAGR of 33.5% is forecasted for the market to reach $20.9 billion by 2032. Further projections indicate a 7% increase in global GDP with its use. With the rise of new entrants such as Google’s Bard and Anthropic’s Claude, the beneficial impact of this technological revolution needs to be analysed and quantified. Currently, Generative AI is enabling leading universities to save years of expensive and laborious drug discovery.
To modernise their digital systems with Generative AI, a major Indian airline recently announced an investment of $200 million. Its latest incarnation ‘AutoGPT’ is being considered the holy grail of AI. It can improve itself by browsing the internet and retaining longer chains of commands for developing computer applications from scratch, consulting on market research, driving autonomous vehicles, predicting the weather, and even detecting cybersecurity loopholes. Thus, there is little doubt that with their ever-growing prowess, LLMs will positively influence industries and individuals.
With self-learning and self-adaptation capabilities, Generative AI can enable unprecedented levels of automation and innovation across various sectors, transforming our interaction with technology. It may enhance lifestyles by offering intelligent digital assistants, personalised healthcare, and autonomous transportation. This will make life more convenient, efficient, and interconnected, ultimately allowing greater access to resources and services.
LLMs are also being subjected to widespread criticism by educators for their ability to complete homework and classwork assignments of all kinds – quantitative, qualitative, analytical, and creative. The bizarre listing of ChatGPT as a co-author in a medical journal recently has baffled researchers around the globe. There are mixed opinions on how the education sector – both school and higher, should deal with such unprecedented innovation.
In systems where education is synonymous with rote learning, the problem is further compounded. While LLMs can help students to develop ‘drafts’ or enhance engagement, the finding that ChatGPT can pass graduate-level exams poses a serious challenge to the global examination system, especially in the online mode. It could also lead to easy and rampant plagiarism in the field of research and publications, thereby limiting the ability of researchers to read, assimilate, synthesize, analyze, and emerge with useful insights as part of their original contribution.
Several academics also believe that LLMs pose serious ethical concerns, can be biased, and are likely to hinder human creativity by snatching away the ability of expression. Nevertheless, any innovation of this magnitude cannot be wished away. It is here to stay and will eventually become ubiquitous.
The question is how do we deal with it? Should we consider it an eclipse on the realm of education or leverage it to drive personalized learning? Do policymakers need to enforce a ban or make Promethean use of it? Does it create an opportunity for educators to relook at the ways in which curriculum, pedagogy, and assessments are designed? Can it empower students to transition from accumulating information to developing the skill of using information in impactful and innovative ways?
According to Global Market Insights, AI in Education has an estimated $4 billion market size, and is expected to grow to $30 billion by 2032. A McKinsey Global Institute Report found that existing technologies could automate 20-40% of activities educators perform. At the same time, technology in the form of online AI tutors has shown 2-2.5 times higher learning outcomes. Hence, it is evident that LLMs may offer abundant prospects to enhance quality of teaching-learning.
For a long time, our education system has been hinged on information acquisition and memorisation. Instead of being passive recipients of the curriculum, students may now have an opportunity to participate in the development of the knowledge economy. It can present insightful ways to modify traditional evaluation methods and incorporate various techniques to assess students’ imagination, knowledge, and skills. Through pedagogical clarity, educators can strategically leverage LLMs as a revolutionary tool.
If incorporated into the curricula, it may facilitate the creation of customised content and immersive simulations. We ought to treat LLMs as calculators were treated for mathematics, rather than treating it as a ‘Man vs. Machine’ problem. After all, education aims at making students creators, and not a replica of computers.
In the years ahead, LLMs may play a vital role as we embrace Pervasive Computing and move towards Society 5.0. With its current trajectory, the hazy picture will soon become clearer. In the meantime, it is essential to wisely apply this cognitive automation to support human growth and socioeconomic development. There will be challenges and we may have to indulge in balancing acts. Yet, such a reality should be translated into Responsible AI to empower academia to nurture future generations.
This content is authored by Dr. Swapnil Morande, Senior Associate, NITI Aayog, and Dr. Shashank Shah, Senior Specialist, NITI Aayog. Views expressed are personal.
