Only a few years ago, the discussion primarily focused on robots, not chatbots. In 2013, Carl Benedikt Frey and Michael Osborne of Oxford University published a report claiming that automation could displace nearly half of all jobs within one to two decades. Their paper created a buzz around automation and its potential economic effects.
Investigative reporters explored the economy, searching for instances of occupations disappearing. Tales were told about automated technology transforming mail delivery, hamburger preparation, harvesting of strawberries, manufacturing of T-shirts, and reporting. Unfortunately, most companies claiming to do this quickly failed.
It has been 10 years since Frey and Osborne predicted automation would cause mass unemployment. However, this has not been the case. Instead, labor markets have become quite tight, and productivity growth rates have slowed significantly, leading to what economists refer to as secular stagnation.
Researchers from OpenAI, the creator of ChatGPT, recently published a working paper asserting that the technology is poised to replace large numbers of jobs. This coincided with the launch of GPT-4 by OpenAI. Will this new wave of robot-based automation prove successful where prior attempts failed?
Once again, the optimism surrounding this new technology is questionable. The paper claims a revolution in AI-powered automation utilizes the same assessment methods as Frey and Osborne. To evaluate how likely it is for tasks to be automated, human experts and ChatGPT were surveyed. The researchers then employed O-NET – an online database of job descriptions – to forecast which professions are prone to automation.
Get Quality Content From Our Trusted Partners
OpenAI researchers have determined that large language models are best suited for various jobs, such as those of central bank monetary authorities and clergy members who require a personal touch. They can also be applied to science, engineering, and sociology.
Frey and Osborne’s 2013 study employed a method adapted from attempts to determine the chances of jobs being done by people overseas for less remuneration. However, this technique did not prove effective in anticipating the susceptibility of occupations to automation and drastically overestimated it.
It is partly because computer specialists cannot accurately foresee computers’ abilities for independent activity. Additionally, the scope of duties engaged with most occupations is more extensive than what O-NET indicates. Occupations such as a teacher or lathe operators differ in their implementation across workplaces in America and even more so when looking at Germany, India, and China.
All aspects of the job market, such as legal regulations, collective bargaining arrangements, wages, competitive strengths, and corporate plans, affect how jobs are altered regarding the technology used and duties expected.
During the 2010s, Volkswagen put much effort into introducing robots to their car assembly lines to enhance production. On the other hand, the most effective vehicle manufacturer on Earth, Toyota, disposed of robots from their assembly lines during that period to make their factories more responsive to changing market trends.
In 2016, the OECD undertook a more accurate analysis of various job types according to Frey and Osborne’s figures. The results suggested that fewer than 10 percent of US jobs were liable to automation – and this number has since been demonstrated as an overestimation. This is likely also true for the statistics featured in the OpenAI researchers’ working paper.
It is suggested that, with the aid of advanced technologies, nearly half of the jobs may be at risk due to large language models. However, further investigation could decrease this number to 10 percent or even less.
The speed of technological developments has not overburdened our economy, though this does not mean it is in a good state. To illustrate how much the job market has changed since 1940, 60% of the roles that people have today didn’t exist back then.
Though it is unlikely that most jobs will be lost and many new jobs may be created due to the introduction of technologies such as ChatGPT, how people work will change. We must modify our outlook on how this alteration takes place.
By the end of the 20th century, a few jobs had been completely replaced by automation. Occupations such as lift operators, film projectionists, and travel agents no longer exist. Although it is difficult to name more than a few, most positions were not wiped out with technological advancement. Instead, the tasks performed within these roles altered; for example, dockers used to load goods manually, whereas now they use cranes.
The skills needed for the same job in the film production industry can vary depending on which country it is in; for instance, Hollywood or Bollywood. Additionally, workers in Sweden typically have more independence when deciding how to do their tasks than those employed in the United States and the UK.
The fact that job changes vary from country to country implies that nothing is predetermined. We must not restrict technological advancements; rather, we should discover new ways for them to be utilized to benefit society instead of causing harm.
With ChatGPT and similar technology, jobs will likely undergo a significant transformation. However, such big changes usually do not happen overnight; they take time to realize them fully. Thus, one should be cautious when presented with reports from companies claiming revolutionary technological progress. It would be wise to wait for actual results before drawing any conclusions.
It is uncertain if software such as ChatGPT, which aids in coding, will either increase or reduce the requirement for developers. Though computer programming, technical writing, and legal writing have been experiencing low productivity growth levels, a decrease in programming costs could result in unmet demand for programmers’ services.
ChatGPT’s large language models have certain drawbacks when used in the larger economy. These AI systems often “hallucinate” and give wrong answers when they cannot answer queries. Gary Marcus, an AI specialist, has said that this is not just a minor issue but a technology limitation that will stay even if ChatGPT’s data set increases.
Chatbots have the potential to offer people a much more sophisticated virtual assistant than Apple’s Siri. This type of AI-enabled companion could read emails, organize meetings and events, write messages, explain concepts we don’t understand, and remind us when something important is coming up. Such an advancement would make a meaningful difference in people’s lives.
The proliferation of digital technologies, particularly social media, and smartphones, has overwhelmed us with data. This inundation can hurt our productivity and mental health. As a solution to this problem, tools should be put in place to help limit the daily disruptions we experience.
Only time will tell if ChatGPT can help or worsen the problems caused by too much information, as it can quickly spread facts and false news. Therefore, these technologies must not be solely left in the hands of Silicon Valley to decide their fate.
ChatGPT is revolutionizing the way we use communication technology, offering unprecedented capabilities for enhancing interactions, automating tasks, and transforming various industries. By leveraging the potential of ChatGPT responsibly and ethically, we can unlock its full potential and shape the future of communication positively and inclusively. As ChatGPT continues to evolve, it holds great promise in shaping how we communicate and collaborate in the digital era, opening up new possibilities for innovation and creativity.
Source: New Statesman
