Introduction

Whether it be used to unlock smartphones, drive vehicles, or write last-minute research papers, artificial intelligence has emerged as a powerful tool for complex operations. In particular, Large Language Models such as ChatGPT have gained a great deal of attention in recent months. These systems are a type of AI that can simulate human intelligence by using internet-based resources and other inputs. As AI becomes more accessible, many anticipate transformative changes in the socioeconomic landscape. 

The Major Players

While more and more players continue to enter the large language model (LLM) market, several large companies have already cemented their place at the forefront of the generative AI field. 

OpenAI: Growing from a small startup to a multi-billion dollar corporation, OpenAI has skyrocketed in popularity with the advent of artificial intelligence technologies. Developing various iterations of the Generative Pre-trained Transformer (GPT) large language models (such as GPT-3 and GPT-4) and the popular chatbot like ChatGPT, the company sits at the cutting-edge of AI innovation. Apart from its work developing the GPT series of LLMs, the company has also built an AI image generator called DALL-E 2 (the successor to DALL-E), leveraging the power of generative AI to reach economic success. The company has successfully attracted billions in investments from Microsoft (another key player in the future of AI) and is on track for rapid expansion. 

Microsoft: Microsoft currently plays a major role in AI development, focusing on both technological innovation and major investments to further the future of the industry. The company aims to integrate AI across its popular Microsoft Office suite while also developing Azure AI to help developers and data scientists “do more with less.” In addition, Microsoft currently partners with OpenAI (with Azure being the exclusive cloud provider of ChatGPT) and has invested over 10 billion dollars in the company to ensure its success, while simultaneously integrating OpenAI’s GPT-4 technology into Bing Search. With the corporation taking a development-investment fusion approach to AI development, Microsoft sits at a uniquely pivotal position in the artificial intelligence space and is set to play a major role in determining the technologies that will shape the coming decade. 

Google: Google has been developing artificial intelligence for over a decade, positioning it as the center of its business strategy over the last few years. After acquiring DeepMind in 2014, the corporation began accelerating its AI strategy, integrating artificial intelligence into both its software and hardware solutions. In addition, the company is currently using LLM frameworks like LaMDA and, more recently, PaLM, and has come out with a generative AI chatbot called Bard just this past year. The company is working on further integrating AI into its existing software (most notably Google Workspace), with experimental programs like Google Labs helping provide users early access to Google’s cutting-edge AI technologies. 

Replacing jobs

In 2022, access to remote work, high inflation rates, and abundant stimulus checks generated a labor shortage. The ‘Great Resignation, a period in which a record 50.5 million workers quit their job undoubtedly harmed the labor market .’ While most of these employees did not leave the workforce altogether, there are currently 3.5 million more job openings than job seekers. Despite economists' predictions that this shortage will persist, the economy appeared reluctant to acknowledge this reality at first. In sectors like health care and manufacturing, output reduced substantially, which in turn reduced wages, creating a continual cycle of economic decline.  For instance, 34% of nurses said in a poll that they would leave their jobs by the end of 2022. More holistically, in the last 50 years, manufacturing’s share of US GDP has decreased by 15%. Until the rapid increase in LLM innovation, a solution for the lack of human resources seemed out of reach. Complex tasks including copywriting, customer service, legal drafting, and content creation can all be solved easily by LLMs. These models can do tasks at record speeds while also eliminating human error. According to a 2022 IBM report, labor shortages have caused 25% of companies to adopt AI as a viable alternative. These adoptions will likely increase productivity (as projected by 64% of business owners) and thus increase GDP by 21% by 2030. Not only will this solve the labor shortage, this growth will create an entirely new job market to ensure that LLMs will only assist the workforce as opposed to replacing it. Indeed, the World Economic Forum predicted that 97 million AI-related jobs will open as a result of AI, opposing worries of workforce displacement. However, given the effectiveness of this technology, it is cost-effective for a business to implement AI as an alternative to paying a human being. Accountants, financial analysts, and other similar professions are at the highest risk for replacement in the near future. 

Yet if a balance is not found between the number of jobs created and the number of jobs lost, the global economy could suddenly collapse. Chinese companies are leading the global race with 58% of them adopting AI in some form. Ironically enough, this tipped the balance in the other direction with the potential to displace 26% of existing jobs in the highly populous country. Iceland, Switzerland, and Estonia, who also attempted to quell the unemployment problem by using AI, are following a similar pattern. If no action is taken, the same phenomenon may occur in the US as AI continues to grow at an annual rate of 37.3%. Overall, the spike in demand for AI will force individuals to adapt while testing the long-term stability of the job market.

Technical Flaws

Despite their immense potential to improve productivity, the output generated by LLMs cannot always be trusted. Users have found that generative AI models like OpenAI’s ChatGPT and Google’s Bard often produce information that does not exist, a phenomenon known as a hallucination. 

Without going into the technological specifics of why hallucinations occur, it is thus imperative to understand that LLMs cannot be relied on for research purposes — all information generated must be cross-verified to make sure that it is accurate. Despite this major flaw, however, LLMs can still be used for the vast majority of cases as companies continue to optimize their models to reduce these unwanted occurrences. 

Ethical Concerns

Companies like OpenAI and Google have faced lawsuits alleging the illegal scraping and reproduction of personal data from the internet. Indeed, a proposed class-action lawsuit filed by Clarkson Law Firm alleged that OpenAI scraped “essentially every piece of data exchanged on the internet it could take” without consent or “just compensation.” 

Similarly, the same firm proposed another class-action lawsuit against Google, Alphabet (its parent company), and DeepMind (its AI subsidiary), claiming that Google “has been secretly stealing everything ever created … on the internet,” including “creative and copywritten works” to build its AI models like PaLM and Bard. 

Seeking a temporary freeze on the commercial access and development of these corporations’ generative AI tools and financial compensation for those affected, the results of this legal proceeding could bear major implications on the future of machine learning and artificial intelligence. 

Conclusion

Similar to the iPhone and the internet, generative AI and large language models in particular are groundbreaking innovations with many potential benefits and harms. It is impossible to foresee the effect of LLMs on the long-term social environment. However, economists have been able to make predictions of worst-case scenarios that should be taken into account when using these tools. AI raises Issues concerning ethicality, fairness, and displacement. Developing robust guidelines for the use of these tools will be critical to ensure responsible AI development in the future.