The software outsourcing industry is one that has always been impacted by developments in Artificial Intelligence. ChatGPT is an AI product launched by the American Company Open AI. It was officially launched as a prototype on November 30, 2022, approximately 59 days ago so it is still a very new product. The current version of ChatGPT is fine-tuned from a model in the GPT-3.5 series, which finished training in early 2022. ChatGPT and GPT 3.5 were trained on an Azure AI supercomputing infrastructure. However there are strong worries throughout the software outsourcing industry that new AI products such as Chat GPT will damage or diminish the software outsourcing industry. Will Chat GPT completely eliminate the need for human labour in the software outsourcing industry? Or will the scenario be slightly different- with Chat GPT simply changing the working environment in the software outsourcing industry. This blog post attempts to answer these important questions.
The chance is, if you are currently reading this blog post, then you have probably heard of and read about Chat GPT and perhaps even know what it does and how it can be used. You may have even created a free account yourself on the website of OpenAI and have been playing around with ChatGPT to your heart’s content. However, for those that are coming across Chat GPT for the first time, ChatGPT is an Artificial Intelligence based chatbot developed by an American company known as OpenAI. This application acts as a sort of two-way chat with users asking a question and the chat interface generating an answer in response to this question. ChatGPT is a language model which can generate a wide variety of text in response to specific questions. The key is to frame the question in such a manner as to elicit a specific answer.
ChatGPT is one of the hot topics in the tech industry all around the world. It has drawn huge admirers as well as critiques just like any new piece of technology. It has its uses as well as its limitations and understanding these limitations and strengths is a key focal point of this article. However, the more pressing purpose of this article is whether and how ChatGPT can affect current jobs in the software outsourcing industry.
Every single individual following the tech industry has heard of ChatGPT and you must have heard of all the disucssion going on everywhere about how this futuristic AI is going to be the next breakthrough in the technology industry. Some analysts are going as far to describe ChatGPT as a technology breakthrough on par with the introduction of the iPhone or the launch of the platform Facebook. The talk in the industry is that Chat GPT is going to replace traditional jobs in the software outsourcing industry.
There is a debate in the software outsourcing industry about the impact of ChatGPT in the short-run and the long-run on jobs and employment in the software outsourcing industry. There are strong arguments on both sides of the debate. One one point of view, ChatGPT and other related AI generative systems will have a significant impact on jobs and employment in the software outsourcing industry, leading to tech layoffs both in the short-run and the long-run. On another point of view, Chat GPT will not have a shock-like effect on the software outsourcing industry and little or no tech companies will layoff their workforce in response to the emergence of ChatGPT. There is also a more nuanced view, which is that the loss of Jobs as a result of ChatGPT may not be felt in the short-run but rather in the longer run.
This idea is creating ripples of fear throughout the software outsourcing industry among employees who feel the impact of the chat programme on their livelihoods, employment and job prospects. To develop this idea further, one must first examine precisely how ChatGPT could replace jobs in the software outsourcing industry. The thinking that prompts this point of view is that jobs and roles currently existing in the software outsourcing industry and performed by humans could be replaced or substituted by an AI generative model such as ChatGPT. The thinking is that big tech companies or even medium-sized companies will use ChatGPT as an excuse to lay off workers and to cut or reduce costs. This is a compelling point and one that needs to be explored further. The other side of the coin is that ChatGPT will not act as a replacement to the need for a human workforce. The idea is that ChatGPT will perhaps replace low-skill or some medium-skilled jobs in the software outsourcing industry. However, complex, high-skilled jobs in the tech sector cannot be performed by a ChatGPT-type application. There are concerns in the market about the accuracy and the precision of the text that is produced by ChatGPT. These points might serve to allay fears among those in the tech industry that the ChatGPT application is not sophisticated enough to replace high-skilled, highly trained labour.
To understand precisely how Chat GPT could impact jobs and employment in the software outsourcing industry we can use the thought experiment of the three boxes. Imagine that there are three boxes- Box A, Box B and Box C. Box A represents very low skilled labour. Box B represents medium skilled labour. Box C represents high skilled labour. These three boxes are in the hands of the employer which for this purpose let us assume is a technology company by the name of “X” which specialises in software outsourcing. Pre or before the introduction of ChatGPT all these three boxes are full. Suddenly, OpenAI announces the launch of ChatGPT creating excitement among technology companies who are anxious to see how it can use this new AI product to improve the performance of their company in software outsourcing. The company makes its decisions box by box. It first turns to Box A which is the most vulnerable of the three boxes. The company leadership decides that the work of 70% (this figure is for explanatory purposes only) of the work and the workforce in this box can be replaced by the ChatGPT function in both the short-run and the long-run.So on this basis, 70% of the labour force in this box is layed off. The company then turns to Box B- the medium skilled labour in the company. It decides that ChatGPT could replace some but not all of the jobs in this box and decides to lay off 45% (this figure is only for explanatory purposes) of labour in this box in the short-run and long-run. The company then looks into the third box, Box C. This is the Box that is most secure and least vulnerable to technological breakthroughs such as ChatGPT. The company weighs its options and decides to not layoff a single worker in this box. Thus according to this analysis highly skilled labour which could include labour with technical skills in the software outsourcing field will be protected from ChatGPT. In short, from an employees perspective, you need to aim to be in Box C which is where you will not be vulnerable to layoffs due to the emergence of applications such as ChatGPT.
This position is based on the point that the ChatGPT application has several inherent limitations. These limitations which are of various types mean that overall ChatGPT cannot act as a suitable replacement for human labour in the software outsourcing industry. To better understand this point one must first look into the key limitations of the ChatGPT programme.
1. Type of content: ChatGPT is an AI learning language model. This means it can only produce text and nothing else. Images, pictures, videos and other material cannot be produced by ChatGPT. This in itself a significant limitation or a limit to the AI Chat programme. ChatGPT can be observed to be purely one-dimensional with the capacity to produce and synthesise text based on its bank of knowledge and training.
2. Training method: ChatGPT has been trained by OpenAI in a specific format. That is to answer user questions. The chat is set up with a user interface. A user can then ask a question and this prompts the chat to answer it. The training that Chat GPT has been created on allows it to draw on its bank of knowledge and information and present them in an accessible format to the user. This means that inevitably ChatGPT reproduces the knowledge it has rather than doing anything else. The analytical or critical reflection element is very limited in ChatGPT. Thus we cannot consider ChatGPT as some sort of ‘brain’, but simply as a useful starting point to acquire information on a topic. The algorithms that have been used to develop Chat GPT are limited in this manner and it is extremely important for users to understand this.
3. Limitations in text creation: One of the primary ways in which Chat GPT can be used is to generate and write text. This can be used in a variety of ways. You can ask ChatGPT to write an essay on a particular topic, or write a poem in a particular manner. You can also use the chat programme to generate research material on a wide variety of topics. The problem here is two-fold. Firstly, there is an accuracy problem. Chat GPT has been known to produce information which is factually incorrect and this is highly problematic. The accuracy problem varies from time to time and depends on the type of text that is being produced. Inevitably some text will be more accurate than others. Thus using ChatGPT to generate text without further fact-checking could prove to be highly problematic as the text and data the chat produces could be inaccurate and this requires further fact-checking on the part of the user. Thus we can conclude that ChatGPT is a useful starting point for generating text but its findings need to be submitted to further rigorous fact-checking.
4. Format: The ChatGPT interface is designed in a particular format- that is a question and answer format. This means that the user has to ask a particular question and the interface will reply with a response to that question. This system means that it is crucial that users ask the correct questions and that they are extremely specific and precise in their instructions. A small change in the wording of the question can lead to a dramatically different answer. Furthermore this question and answer format has its own inherent limitations which is that it is more complex, cumbersome and less user-friendly when compared to a search engine such as Google or Bing which produces a large volume of information in response to the typing of some search terms.
5.Technical assistant: Chat GPT can be used to produce technical data or information which can be of use to software developers, coders and programmers, software engineers, prompt engineers and solutions engineers. The type of content that ChatGPT can produce is varied. For instance a user can ask ChatGPT to write the code for a simple computer programs using the language C++ or C#. Furthermore, it can then ask ChatGPT to write the code for a basic algorithm using the C++ programming language. It can also request the Chat to develop this algorithm further and write a more advanced or complex algorithm using the same C++ language. Furthermore ChatGPT can be used to write the code for a basic HTML website and then be used to write more advanced code. In this way the chat can be used to generate several lines of code. This makes the chat an ideal programming assistant or facilitator. However, there is a problem here too. Chat GPT is limited in the nature of the code or algorithms it is able to produce. It cannot produce the complete line of code required and is unable to produce highly advanced lines of code of algorithms. This makes ChatGPT a useful programming assistant for programmers but nothing more.
So what does all the limitations outlined above mean for jobs in the software outsourcing industry? Well, the answer is quite simple- there is no need to panic, at least not yet. ChatGPT cannot and will not replace jobs in the software outsourcing industry. There is little need to debate as to whether jobs will be lost in the high skilled sector but an argument can be made that jobs will not be affected even in the low-skilled and medium-skilled sector. This is for three reasons. Firstly, ChatGPT is overall too simplistic as outlined above. Thus it cannot replace or act as an effective substitute for jobs and the skills provided by human labour in the software outsourcing industry. Secondly, ChatGPT is filled with inaccuracies in the data, information and text it produced. Echoes of inaccuracies have become more than shadow or whisper and have to be looked at seriously. The third reason is that AI generative applications such as ChatGPT can not be sufficiently technical to replace the technical roles in the software outsourcing industry. ChatGPT according to the observations made above can be said to not be able to meet the technical requirements of job roles in the software outsourcing industry. The chat programme is too superficial and merely scratches the surface. On the whole therefore, workers in the software outsourcing industry need not be concerned about losing their jobs to ChatGPT and other AI generative models. As outlined earlier, much of this depends on the ‘Box’ the particular labourer is in- and low-skilled labour will be more vulnerable than high skilled labour. Furthermore there is also a need to distinguish between job losses in the short-run and long-run.
There are five separate points that can be made to illustrate how ChatGPT can affect jobs in the software outsourcing industry. The first point is the speed point. ChatGPT operates at remarkable speed, generating text, code, research and other material in an extremely short period of time. It would be extremely difficult for low-skilled labour in particular to match this speed of operation which would make ChatGPT a suitable replacement for low-skilled labour (Box A). However, the position in relation to medium skilled labour (Box B) and high-skilled labour (Box C) would be different. ChatGPT would find it more challenging (but not impossible) to replace jobs in this sector as labour in this sector can also operate at a very high speed. Thus, whether ChatGPT can replace jobs in the low skilled, medium skilled and high skilled boxes is a contestable point which is open to debate and discussion. On balance though, it seems possible that due to the speed or operation of ChatGPT and the efficiency of the application that ChatGPT will be able to replace some jobs in the software outsourcing industry, leading to tech layoffs. The second point is about knowledge and information. ChatGPT has an almost infinite bank of knowledge, information, data and statistics in its bank from a variety of sources. Not a single human labourer will be able to match this level of knowledge storage, including in Box C, the high-skilled labour box. It is this second point that is very concerning as the sheer volume and quantity of knowledge and information ChatGPT can produce is so vast (regardless of accuracy problems discussed earlier in this blog post) that ChatGPT could easily outstrip a human competitor. From this perspective, ChatGPT could make all three Boxes -A, B and C, low-skilled labour, medium-skilled labour and high-skilled labour vulnerable due to the fact that any individual in either on of these three boxes would not be able to compete with ChatGPT based on knowledge alone. The third point is related to technical ability and skills. ChatGPT clearly is able to produce technical data and information. It can produce snippets of code, write software and algorithms and even write the basic code for websites. Programming and the writing of code is a highly technical field requiring an advanced skill set. ChatGPT could replace basic, low-skilled technical jobs in the software outsourcing industry with ease. However, high-skilled advanced technical jobs in software outsourcing are unlikely to be vulnerable to ChatGPT.
This blog post has analyzed the ChatGPT programme recently introduced by OpenAI from the perspective of the software outsourcing industry. In particular this blog post attempts to understand how ChatGPT could affect jobs and human labour in the software outsourcing industry. To explain this idea a thought experiment of the three boxes was used- Box A, Box B and Box C. Furthermore two positions were considered- that ChatGPT will lead to job losses in the software outsourcing industry and a second position that ChatGPT will not lead to job losses in the software outsourcing industry. This analysis was looked at from the perspective of the short-run and the long-run as well as from the perspective of low-skilled, medium-skilled and high-skilled labour. In conclusion, there is not yet a clear view on the precise impact of Chat GPT on the software outsourcing industry. There are arguments and points of view on both sides. But one point is clear, regardless of which side of the debate you are on, AI generative tools such as ChatGPT will have a profound impact on the software outsourcing industry throughout the days, weeks and months to come.
Note: The above blog post is based on the current version of ChatGPT released by OpenAI. The current version of ChatGPT is fine-tuned from a model in the GPT-3.5 series, which finished training in early 2022. ChatGPT and GPT 3.5 were trained on an Azure AI supercomputing infrastructure. If and when a subsequent version or model of ChatGPT is released by Open AI the arguments, positions, thoughts and conclusions in this blog post could change. EFutures reserves the right to make changes to the content of this blog post accordingly. This blog post and its writings, thoughts, arguments, thought experiments, analysis, content and other indicia is the exclusive property of EFutures (Pvt) (Ltd) and is protected by copyright, trademark and Intellectual Property Law. No part of this blog post including text, arguments, ideas or analysis may be reproduced in any form without the prior written consent of Efutures (Pvt) (Ltd). This blog post was prepared by the content creation team at Efutures (Pvt) (Ltd).
Stay updated with our tech articles – your go-to source for the latest insights, trends, and innovations in technology.