(Second of two parts)
When ChatGPT 3.5 was released last year, it made global headlines for its ability to perform tasks such as analyzing professional contracts and complex spreadsheets. ChatGPT is a rapidly evolving text-based artificial intelligence (AI) that facilitates “human” interactions via its natural language responses. Despite its nascency, ChatGPT has already solidified its presence in various industries. Its myriad of functions (e.g., content creation, data analysis, and code generation) can help organizations enhance their products and services, streamline work processes, and refine customer service.
However, many are also deeply concerned about its use in business, education and various other sectors. In the first part of this article, we discussed the science behind Generative Pre-Trained Transformer (GPT), hot topics regarding its human aspect, its response biases, and potential business applications. In this second part, we discuss the practical ways ChatGPT can be used in business and the potential risks it presents.
PRACTICAL BUSINESS USE CASES
With the ability of AI to automate several tasks, businesses can reduce their labor costs while simultaneously enhancing workflows. ChatGPT’s flexibility can support and possibly even enhance various corporate functions such as customer service, data analytics, sales and marketing, and finance.
Customer service. Given its text-based nature, ChatGPT can leverage its ability to customize responses based on user prompts to facilitate a seamless user experience. The program’s versatility means that it can be incorporated into different platforms such as chatbots, e-mail, and SMS. ChatGPT can provide round-the-clock support, potentially becoming instrumental in the banking, healthcare and information technology (IT) industries. Small- and medium-sized enterprises (SMEs) can capitalize on AI by setting up a chatbot that can interface with customers without needing human moderation. Since the AI will continue to evolve through repeated customer interactions, the company can make use of the data for continuous improvement.
Data analysis. ChatGPT has a wealth of information to draw on, potentially making it an asset for tasks such as market research, research and development, and financial forecasts. Businesses will be able to analyze data more efficiently given its comprehensive set of information. A practical example would be ChatGPT’s capacity to break down complex code and generate bug fixes.
Sales and marketing. Sales and marketing are corporate functions that require a more personalized approach, and ChatGPT can address this by utilizing its natural language model to create bespoke solutions. Apart from generating SEO-friendly keywords to outlining drafts, ChatGPT can also produce personalized e-mails, blogs and video ideas.
Despite ChatGPT’s potential for streamlining operations, it can pose risks for organizations. Given the nature of this AI and how it can evolve (i.e., it analyzes large data sets on the internet before generating a response based on the user prompt), security, accuracy and fairness are paramount concerns.
A potential pitfall for the AI lies in its primary competency — that it can facilitate more “human-like” interactions since humans are prone to error and subject to different biases. Its very strength can prove to be its weakness, since the conveniences it affords can also facilitate the spread of disinformation, ethical issues and copyright disputes.
Data accuracy. OpenAI, the company behind the program, acknowledged that the software produced erroneous and/or biased content. One of the program’s limitations is that its learning model was programmed in 2021, which means that it has little-to-no knowledge of developments since then. It is also worth mentioning that not all online information is accurate, proving to be a substantial constraint for ChatGPT. People have even claimed that the AI can “hallucinate” because it has populated user queries with false information, such as listing down incorrect credentials for public individuals.
Cybersecurity and data privacy. Its online nature makes ChatGPT vulnerable to cybersecurity attacks that make it a potential risk to businesses. The program can endanger one’s privacy because it can sift through a vast range of data accessible online. Businesses will have to deliberate whether the technology’s benefits outweigh its potential security risks. They must also be vigilant when it comes to the security of both themselves as well as their clients.
Bias. In the first part of the article, we discussed how ChatGPT has a category of answers that consists of subjective responses. This inherent bias may deter corporations from assimilating it into their established work systems. There was a case wherein ChatGPT was asked which airline passengers could pose a risk, and it asserted that individuals who traveled to North Korea, Afghanistan, Iraq and Syria were the more prominent dangers. The learning model is continuously evolving, but it still needs some form of arbitration to avoid ethics and bias-related issues.
Ethics. In academia, there have been longstanding, divergent opinions when it comes to technological advancements, ranging from the archaic decision of whether smartphones should be allowed in class, to more current concerns, such as the ethics of using AI to accomplish assignments and/or examinations. The jury is still out as institutions have varied responses, with some universities mandating the return of in-person exams to safeguard against cheating, whereas others have started to delineate AI-specific guidelines. Plagiarism, however, remains a principal concern. The convenience of AI may exacerbate issues when it comes to the originality of work, whether in academia or corporations.
Intellectual property and copyright. In light of ongoing discussions that ChatGPT can replace, or at the very least, assist with certain types of work, it is vital to understand the legal repercussions. With copyright protection, the US Copyright Office will not register work that was generated by an AI. In accordance with US law, AI-generated output will either be a claimless work available in the public domain or considered a derivative work of the tools that the AI was developed upon. This raises the question as to who the true owner is — the creators of ChatGPT or the user for whom the output was generated.
AI IS HERE TO STAY
As one of many developing AIs, ChatGPT offers advantages and risks for personal users and organizations. It is also apparent that human intervention is necessary to truly leverage its benefits and mitigate its intrinsic shortcomings.
With the current technological climate though, it seems that businesses will no longer be able to turn a blind eye to the program. Similarly, OpenAI is not the only company making headlines when it comes to artificial intelligence as other companies are racing to develop their own versions. One thing is clear — AI is here to stay.
Technology is at the forefront of business change and learning how to leverage it is critical. ChatGPT has jolted the corporate landscape, presenting both challenges and opportunities for organizations. For companies considering the use of AI, it is vital to evaluate its role in their respective ecosystems. Ultimately, the biggest question is whether hypothetical returns will be enough to mitigate the potential risks.
This article is for general information only and is not a substitute for professional advice where the facts and circumstances warrant. The views and opinions expressed above are those of the author and do not necessarily represent the views of SGV & Co.
Randall C. Antonio is a technology consulting partner of SGV & Co.