"an open white toolbox with tools and tentacles coming out from it, white background", image generated with MidJourney
In artificial intelligence domain, natural language processing (NLP) is a prominent subject in information technologies. F. Scarcello describes it in the Encyclopedia of Bioinformatics and Computational Biology (2019) as follows: “Providing machines with the ability to communicate by written and spoken natural languages is one of the first and most widely studied objectives in AI and computational linguistics.”
Research on NLP has been foundational to generative AI (Gen-AI), from the writing capabilities enabled by large language models (LLMs)1 to the text-to-image tools, as explained by IBM. Among the multitude of products and services based on NLP, one of the most famous is chatGPT from OpenAI. Launched in November 2022, it has gained 1.8 billion users in ten months and triggered global reactions, from a state ban in Italy to a writer’s union strike in the US. ChatGPT, and more generally, Gen-AIs like Microsoft’s Copilot or MidJourney, symbolize the global digital transformation of society. Free, online, and easy to use, these services impact all sectors of activity, whether commercial or institutional, global or local. This article describes some effects of these new digital tools on the information systems of businesses and institutions by referring to the “VUCA” leadership theory. This is a topic of interest as stated by Getchell et al. (2022), who propose a research agenda on the matter.
Volatility refers to the speed and unpredictability of change. While Gen-AI tools accelerate content production, it is unclear if they improve efficiency or just add noise. Grammarly, offering “free AI Writing Assistance”, published a report based on a poll among 1,002 knowledge workers2 and 253 business leaders, showing a gap in AI proficiency and use between leaders and workers. Gen-AI helps leaders to manage volatility by providing insights for faster and better decision-making but also adds unpredictability by radically changing communication processes for workers. In addition, the fast evolution of Gen-AI contributes to volatility as frequent updates and changes in tools’ capabilities can be an issue for businesses to keep up with, not counting the changes in pricing.
Gen-AIs’ apparent efficiency creates dependency and can open new vulnerabilities when misused, producing inaccurate content. It generates answers based on statistical models and past data without guaranteeing correctness or appropriateness, thus creating uncertainty by lacking predictability and making information harder to rely on. Task automation also contributes to uncertainty among employees, raising concerns about job security, as discussed by Manjeet Rege and Hemachandran K. (2024).
Complexity arises from the interdependence of various business environment factors. Adding Gen-Ai to existing systems increases IT infrastructure complexity, requiring specialized skills for integration and maintenance. With customized LLMs, businesses are inclined to create their own Gen-AI from their knowledge base. While having an internal chatGPT is tempting, implementing it properly is costly. Complexity is also linked to data-driven business culture. Having an efficient customized LLM needs structured knowledge bases, which is not the case in SMEs, as this WEF article indicates.
Ambiguity is also a concern. As highlighted above, Gen-AI outputs need human evaluation. However, the authoritative tone and the clarity of the writing often lead to skipping this step. This “ignorance bias” is one problematic effect described by Forbes. Ambiguity comes also with ambiguous and open to multiple interpretations prompts that lead to decision-making based on poorly processed information. Lastly, the complex ethical and regulatory questions make it challenging for entities to define unambiguous guidelines on what constitutes appropriate use of Gen-AIs.
In conclusion, Gen-AIs are powerful tools for enhancing information systems and decision-making. But it comes with downsides that we should be more aware of. Gen-AIs’ “magic” hides dark sides as revealed by Times Magazine about Kenyan employees paid 2$ per hour to “Make ChatGPT Less Toxic”.
Activities General ISDT Module personal thoughts Reflective practices Search and Social Media Marketing
References:
2024 State of business communication. (n.d.). https://go.grammarly.com/thankyou/content/2024-SOBC-Report
Breaking down the cost of large language models | QWAK. (n.d.). https://www.qwak.com/post/llm-cost
Dwivedi, Y. K., Kshetri, N., Hughes, L., Slade, E. L., Jeyaraj, A., Kar, A. K., Baabdullah, A. M., Koohang, A., Raghavan, V., Ahuja, M., Albanna, H., Albashrawi, M. A., Al-Busaidi, A. S., Balakrishnan, J., Barlette, Y., Basu, S., Bose, I., Brooks, L., Buhalis, D., . . . Wright, R. (2023). Opinion Paper: “So what if ChatGPT wrote it?” Multidisciplinary perspectives on opportunities, challenges and implications of generative conversational AI for research, practice and policy. International Journal of Information Management, 71, 102642. https://doi.org/10.1016/j.ijinfomgt.2023.102642
Subject Guides: Using Generative AI: Ethical Considerations. (n.d.). https://guides.library.ualberta.ca/generative-ai/ethics
Getchell, K. M., Carradini, S., Cardon, P. W., Fleischmann, C., Ma, H., Aritz, J., & Stapp, J. (2022). Artificial intelligence in Business Communication: The changing landscape of research and teaching. Business and Professional Communication Quarterly, 85(1), 7–33. https://doi.org/10.1177/23294906221074311
How can SMEs become data-driven enterprises? (2024, June 13). World Economic Forum. https://www.weforum.org/agenda/2023/06/how-can-smes-become-data-driven-enterprises/
Perrigo, B. (2023, January 18). Exclusive: OpenAI used Kenyan workers on less than $2 per hour to make ChatGPT less toxic. TIME. https://time.com/6247678/openai-chatgpt-kenya-workers/
Scarcello, F. (2019). Artificial intelligence. In Elsevier eBooks (pp. 287–293). https://doi.org/10.1016/b978-0-12-809633-8.20326-9
University of St. Thomas. (2024, June 13). Tommie Expert: Generative AI’s Real-World Impact on Job Markets. Newsroom | University of St. Thomas. https://news.stthomas.edu/generative-ais-real-world-impact-on-job-markets/#:~:text=Furthermore%2C%20a%20Forrester%20projection%20indicates,6.9%25%20or%2011.08%20million%20jobs.
What is NLP (Natural Language Processing)? | IBM. (n.d.). https://www.ibm.com/topics/natural-language-processing
Who first originated the term VUCA (Volatility, Uncertainty, Complexity and Ambiguity)? - USAHEC Find Your Answer. (n.d.). https://usawc.libanswers.com/faq/84869
- LLM: Large Language Model. A complex mathematical representation of language that is based on very large amounts of data and allows computers to produce language that seems similar to what a human might say. LLM. (2024). In English Meaning – Cambridge Dictionary. https://dictionary.cambridge.org/dictionary/english/llm ↩︎
- On page 4 of the report, the definition of a “Knowledge Worker” is written: “Employees working full-time at corporations with 150 employees or more.” 2024 State of business communication. (n.d.). Page 4. https://go.grammarly.com/thankyou/content/2024-SOBC-Report ↩︎
Laurent Bolli - Blog 