Chat GPT was introduced to big hype in November 2022 by OpenAI, an American artificial intelligence research lab. Interestingly, Microsoft just invested $10bn in Openai, presumably to be able to use the technology on its platforms. Google is not far behind either with its chatbot called “Apprentice Bard,” and Baidu, the Chinese search engine, plans to add a chatbot to its search engine in March.
Google’s AI chief, Jeff Dean, is said to have told employees that its slower release is due to the fact that the company has much more “reputational risk” in providing wrong information and thus is moving “more conservatively than a small startup.” He is not wrong! The GPT-3 series of big language models from OpenAI provides the foundation for ChatGPT, which has been optimized using supervised and reinforcement learning methods. ChatGPT can produce lines of code, emails, essays, and poems in response to a prompt or question asked.
Read more: Can they snatch our jobs? Robots on the frontline to defeat the coronavirus
The EU Agency for Fundamental Rights has warned that algorithms built on inaccurate data could be harmful and that precautions should be taken to reduce bias and discrimination. Users were forewarned by Meta last year that the Blenderbot 3 chatbot may make erroneous or conflicting claims, forget some facts, and “forget that they are a bot.” Since ChatGPT is trained on statistical correlations and patterns, it does not comprehend input or output as a human would. Due to the lack of transparency surrounding the training of AI tools, it can be challenging to determine how the bot arrived at a conclusion. AI tools can be skewed depending on the data they are trained with.
Concern over ChatGPT has also heightened in recent weeks among educators. Some viewing AI tools as outright cheating, a group of Australian universities announced they would alter exam forms to do away with them. However, others in the educational field are less concerned about using AI technologies in the classroom; some even see this as an opportunity rather than a threat. Where basic facts are AI-derived, and students can focus on offering differentiated higher-level analysis, a skill still not possible with current AI chatbots.
ChatGPT is part of a long-running arms race between AI-authoring tools and programs that try to detect them. An amateur programmer revealed that he had spent his New Year’s holiday break developing an app to analyze texts and determine whether ChatGPT generated them. His app, GPTZero, was not the first in the industry and will not be the last. It takes little imagination to see a future in which each essay is run through an AI detector because universities currently already employ tools that can detect plagiarism. Digital watermarks or other types of signifiers that will identify AI work are also being proposed by campaigners.
Read more: Can & Will Robots render Humans useless?
AI is profoundly changing how humans live. In addition to collecting and processing data infinitely faster than humans, machines that require human-level intelligence may also learn and make their own decisions. As processing power becomes more accessible, AI is becoming mainstream. Science fiction is turning into reality. There are worries that AI will soon rule the world of humans. The late Stephen Hawking, Bill Gates, and even Elon Musk have periodically warned the public about “super intelligence” that will surpass human intelligence. Simple customer service roles, repetitive or routine work that can be automated, data entry and processing, and some types of content production are just a few of the occupations that AI tools like ChatGPT may replace. By 2025, AI and robotics may eliminate 85 million jobs, but 97 million jobs may also be created due to these changes, according to the World Economic Forum’s Future of Jobs Report from 2020. However, not everyone will be affected equally by this interruption; women, younger people, and those making less money will be particularly hard hit, especially in places like Pakistan.
Low-skill manual and high-skill mental tasks are among the most prone to automation. Along with regular low-end professions, many middle-income occupations requiring programmable cognitive abilities, such as accounting, mortgage origination, paralegal work, and back-office transaction processing, will see the biggest decreases in employment.
Three–four million young people in Pakistan use the internet economy and global connectivity to connect with work prospects abroad. The availability of Pakistani employees online has grown dramatically, according to the Oxford Online Labor Index (OLI). Surprisingly, Pakistan provides more than twice as much online freelance labor per capita as India, which has a population of 1.4 billion and contributes around one-third of the world’s online freelance workforce. Pakistan is currently the third-largest contributor to online labor globally in terms of numbers, just behind Bangladesh and ahead of the US. The main problem is constantly assessing how secondary and postsecondary education relates to the shifting skill needs of the international labor market.
Pakistan is likely to become more and more negatively susceptible to advanced AI technologies like ChatGPT since it is providing lower-skilled freelancing work that is likely to get replaced by AI technologies, including meeting transcriptions, basic content writing, graphic designing, and so on.
Dr. Abdul Ghaffar Mughal, an economics lecturer at Northeastern University in Boston, recently wrote in an article for GVS that “the task for educators needs to be redefined in terms of the transmission of needed skills in a fast-changing world to future cohorts while creating greater incentives for students to enroll in STEM as well as vocational training programs.” Pakistan’s labor force must acquire new skill sets; the market need not collapse with the Fourth Industrial Revolution, but rather, it must integrate with the changes by moving away from jobs that AI is replacing and toward those it is creating.