The transformative potential of Generative Artificial Intelligence (GenAI) in business: a text mining analysis on innovation data sources

El potencial transformador de la Inteligencia Artificial Generativa (GenAI) en el ámbito empresarial: un análisis de texto de fuentes de datos de innovación

Enrique Cano-Marin

Computer Science Department, University of Alcala

enrique.canom@uah.es

https://orcid.org/0000-0002-7948-1657

Received: 20-11-2023; Accepted: 07-03-2024; Published: 31-05-2024

Abstract

Objective:

This study investigates the transformative potential of Generative Artificial Intelligence (GenAI) within the business domain and the entrepreneurial activity.

Methodology:

A comprehensive research design is adopted, integrating text-mining techniques to analyse data obtained from publicly available innovation repositories. A systematic literature review (SLR) is developed based on the literature obtained from all databases indexed in Web of Science (WoS), incorporating preprints from arXiv, alongside industry-related innovation data in the form of patents from Google Patents. This method enables the derivation of valuable insights regarding the impact and prospective developments of GenAI across diverse business sectors and industries by leveraging Natural Language Processing (NLP) and network analysis.

Results:

The research outcomes highlight the significant potential of GenAI in enabling informed decision-making, enhancing productivity, and revealing new growth opportunities in the business landscape. The continuously evolving business environment is examined, emphasising GenAI’s role as a catalyst for data-driven innovation. However, there are still relevant limitations to overcome.

Limitations:

The selection of data sources and the study period may have excluded relevant or recently published articles and patents within the scope of the present research. The language of the databases analysed is only English.

Practical Implications:

The practical implications of this study carry significant weight, serving as a valuable resource for decision-makers, researchers, and practitioners navigating the constantly shifting terrain of business innovation through the lens of GenAI. Understanding the potential advantages and challenges associated with GenAI adoption equips stakeholders to make informed decisions and develop future business strategies.

Keywords: Generative Artificial Intelligence (GenAI); Large-Language Models (LLMs); business; innovation; Natural Language Processing (NLP)

JEL Codes: D80; L86; O32

Resumen

Objetivo:

Este estudio investiga el potencial transformador de la Inteligencia Artificial Generativa (GenAI) en el ámbito empresarial y en el emprendimiento.

Metodología:

Se adopta un diseño de investigación que integra técnicas de minería de texto para analizar datos obtenidos de repositorios de innovación públicamente disponibles. Una revisión sistemática de literatura (SLR) se desarrolla abarcando las fuentes de literatura científica procedente Web of Science (WoS), incorporando preprints de arXiv, junto con datos de innovación de Google Patents. Este método facilita la obtención de insights sobre el impacto de GenAI en diversos sectores e industrias empresariales mediante el aprovechamiento del Procesamiento del Lenguaje Natural (NLP) y el análisis de grafos.

Resultados:

Los resultados de la investigación resaltan el potencial de GenAI para facilitar la toma de decisiones informadas, mejorar la productividad y revelar nuevas oportunidades de crecimiento en el panorama empresarial. Se examina el papel de GenAI como catalizador para la innovación impulsada por datos. Sin embargo, aún existen limitaciones por superar.

Limitaciones:

La selección de fuentes de datos y el período de estudio pueden haber excluido artículos y patentes relevantes o de reciente publicación dentro del alcance de la presente investigación. El idioma de los textos analizados es el inglés.

Implicaciones Prácticas:

Las implicaciones prácticas de este estudio sirven como un recurso valioso para líderes, investigadores y profesionales en innovación empresarial a través del uso de la GenAI. Comprender las ventajas potenciales y los riesgos asociados capacita a los interesados para tomar decisiones informadas y definir estrategias empresariales futuras.

Palabras clave: Inteligencia Artificial Generativa (GenAI); Modelos Grandes de Lenguaje (LLMs); negocios; innovación; Procesamiento de Lenguaje Natural (NLP)

Códigos JEL: D80; L86; O32

1. Introduction

Technological progress plays a crucial role in driving innovation, shaping economic growth, and influencing competitive strategies (Akter et al., 2023). The current business landscape is undergoing a significant digital transformation, primarily fuelled by the rapid growth of unstructured data (Elia et al., 2022). As companies deal with vast amounts of information presented in diverse formats, the critical need to use this data for strategic advantage becomes increasingly clear (Zhang et al., 2021). Noteworthy studies reveal that around 80% of the data within a typical company is unstructured (Faccia et al., 2022), meaning that is not adhered to any data model, and emphasising the substantial challenge of managing and extracting value from such a significant portion of the information landscape (Möhring et al., 2022). Aligned with Oesterreich et al. (2022), organisations face challenges in extracting value from their data repositories during the journey towards digital transformation. This struggle revolves around the imperative to adapt both internal operations and external product and service offerings through innovation (Ahamat & Sin, 2022), ensuring competitiveness in the evolving landscape (Tagscherer & Carbon, 2023). This involves investigating advanced technologies capable of extracting valuable insights from unstructured data, ultimately expediting the digital transformation of organisations. This process was significantly accelerated by the impact of COVID-19 (Martínez et al., 2022) and can lead to improved firm performance (Heredia et al., 2022).

Among the technologies driving digital transformation, Artificial Intelligence (AI) plays a pivotal role (Ahmed et al., 2022). It demonstrates a proficiency in deciphering complex patterns, assimilating large data sets, and navigating decision-making processes with agility (Enholm et al., 2022). As organisations deal with the challenges of the digital era, AI’s ability in comprehending, predicting, and adapting to changing contexts, positions it as a leading force in driving transformative technological advancements (Perifanis & Kitsios, 2023).

Building upon this foundation, Generative AI (GenAI) has disrupted the field of Natural Language Processing (NLP) with applications across different industries. GenAI uses sophisticated algorithms to comprehend and interpret unstructured data, showcasing not only the ability to process information, but also to autonomously generate coherent and contextually relevant content (Salah et al., 2023). Going beyond textual information, GenAI extends its capabilities into multimodal applications, seamlessly integrating with diverse data types. This adaptability enables it to contribute to a variety of fields, including software engineering (Daun & Brings, 2023) or drug development (Zhao & Wu, 2023), among others. With this versatile approach, GenAI stands out across industries as a potent tool for businesses that seek effective data management and the extraction of actionable intelligence from it. This translates into the development of new products, services, and capabilities, ultimately enhancing efficiency (Kanbach et al., 2023) and business innovation development (Mariani & Dwivedi, 2024).

Above all GenAI tools, ChatGPT stands out. It is a chat-based application that leverages language models developed by OpenAI based on the GPT (Generative Pre-trained Transformer) architecture (OpenAI, 2022). It was launched on November 30th 2022 and reached one million users in less than five days after its launch, and it is causing significant disruption in different fields (Dwivedi et al., 2023b).

However, despite the potential benefits that GenAI offers, a discernible gap exists in understanding its optimal integration within business frameworks. As organisations navigate the intricacies of adopting GenAI, questions arise regarding its ethical implications, data security considerations, and the overarching impact on existing workflows (Ray, 2023). This paper seeks to address this gap by critically examining the usage of GenAI in businesses and entrepreneurship, shedding light on both its transformative potential, the existing opportunities and potential challenges that warrant careful consideration in the era of digital evolution. To do so, the study consists of an analysis of innovation data sources, encompassing Google Patents for patents and Web of Science for academic publications. This inherently includes arXiv covering preprints in the scope, given the novelty of the topic.

The paper follows the next structure: Section 2 includes a comprehensive literature review that contextualises the study, followed by an explanation of the chosen methodology in Section 3. Subsequently, Section 4 presents the results and discusses the findings. In section 5, the main challenges in the implementation of GenAI in businesses are presented. Finally, Section 6 is a concluding section that summarises the key takeaways, including the practical implications, an examination of the limitations encountered during the study and potential avenues for future research.

2. Literature Review

2.1. The disruptive potential of GenAI

The rapid evolution of AI has brought forth transformative advances, reshaping the landscape of various industries. GenAI is a branch of AI that is capable of generating content based on generative models, and offers unprecedented capabilities in the synthesis and creation of diverse forms of content. Based on the transformer architecture proposed by Vaswani et al. (2017), GPT is an OpenAI architecture launched in 2018 that uses artificial neural networks pre-trained on large datasets of unlabelled text. These are called Large-Language Models (LLMs) and are capable of generating novel, human-like content and learning the intricacies of input data, enabling them to produce coherent and contextually relevant outputs. GenAI has disrupted the NLP field and the analysis of unstructured textual information. GenAI has demonstrated capabilities in understanding context, semantics, and linguistic nuances that can be applied in machine translation, question answering (QA), Named-Entity Recognition (NER), data cleansing and preprocessing, text summarisation and sentiment analysis (Wang et al., 2023b).

Unlike traditional AI systems that follow predefined rules, GenAI models possess the remarkable ability to autonomously generate and understand content across multiple modalities, if trained on data of different formats. This inherent multimodality empowers businesses to harness the potential of AI, not only in textual domains, but also in the generation and analysis of images, videos and audios (Zhao et al., 2023b). In the case of GenAI and image generation, Handa et al. (2023) explored the role of ChatGPT in medical imaging and its benefits. These models could be trained also on complex structures, like protein sequences, which could be used in drug discovery (Madani et al., 2023).

2.2. GenAI in businesses

ChatGPT has effectively lowered the entry barrier for the use of advanced state-of-the-art LLMs (like GPT 3.5 or GPT 4) and, with over 100 million weekly active users, it stands as the fastest-growing consumer application to date (Eysenbach, 2023). In addition, Noy & Zhang (2023) have investigated the effects of ChatGPT on enhancing workers’ productivity, concluding that it has a positive impact in terms of job satisfaction and self-efficacy. Among the applications derived from GenAI, the usage in software engineering is highlighted, and the implications of GenAI in software engineering span beyond code generation. It encompasses areas such as data science, bug detection, code optimisation, predicting potential security vulnerabilities and assisting developers in creating more robust and reliable software (Hassani & Silva, 2023). In the field of healthcare, authors like Cascella et al. (2023), Sallam (2023), Dave et al. (2023) and Vaishya et al. (2023) explored the use of generative AI in medicine, highlighting how tools like ChatGPT could assist healthcare professionals in clinical and laboratory diagnosis, learning processes and literature review, and could support patients with virtual assistants in the development of personalised medicine. However, they all agreed that there are existing limitations (e.g. hallucinations): the ethical concerns derived from its usage, which should be thoroughly considered, and the need for regulation, aligned with Meskó & Topol (2023). In education, ChatGPT and, in extension, LLMs, could be used for content generation, learning experience personalisation and improving student engagement, as long as fact-checking, bias mitigation mechanisms and guidelines are in place (Kasneci et al., 2023). Nevertheless, as outlined by Cotton et al. (2023) and Perkins (2023), concerns emerge regarding academic integrity and plagiarism, emphasising the imperative for fostering critical thinking, problem-solving, and communication skills. In addition, Dwivedi et al. (2023b) assessed the potential of GenAI across various fields. They underscored its possible applications in marketing, aiding in content generation, market research, and the development of marketing campaigns.

In summary, the general-purpose character of GenAI makes this technology a new foundation to be leveraged by any kind of business.

Thus, the following research questions are posed:

RQ1: How is Generative AI innovation manifested in practical applications across diverse domains?
RQ2: How do research studies investigate the incorporation of Generative AI in fostering business innovation and strategies?
RQ3: How do patents mirror the impact and adoption of Generative AI in various industries?
RQ4: What risks and opportunities are linked to the implementation of Generative AI in business operations and strategies?

3. Methodology

In this research, a systematic literature review (SLR) has been completed in combination with a patent analysis to reflect the challenges and opportunities in the innovation and adoption of GenAI in businesses. It was adhered to the SPAR-4-SLR framework (Paul et al., 2021), employing a transparently reproducible, end-to-end process, consisting of three distinct phases: assembling, arranging, and assessing. During the assembling phase, literature was gathered and consolidated from diverse sources, including patents. In the arranging stage, the literature underwent systematic processing and organisation, and texts were pre-processed. The assessing phase involved the insights generated from the literature within the study’s scope, leveraging text mining techniques. Figure 1 illustrates said methodology.

3.1. Data collection, sample and preprocessing

Starting with the assembling phase of the SLR, Figure 1 shows the search procedure used to carry out the research. 887 scientific articles in scope were collected from Web of Science (WoS), including all databases indexed into this relevant academic metasearch engine (Martín-Martín et al., 2021). The analysis includes preprints from arXiv.com, a preprint server for physics, computer science, and related fields (Breuer et al., 2022). In addition, 409 patents were collected from Google Patents, which indexes over 20 million patents (grants and applications) worldwide (Noruzi & Abdekhoda, 2014). In the need for innovation to gain competitive advantage, companies explore sources of knowledge that could be leveraged to enhance their firm’s performance. Aligned with Guo et al. (2022), patents provide a mean to learn from competitors. For this reason, Google Patents is used as a source to gauge the level of activity in the industry, regarding the development and adoption of GenAI systems.

In order to reflect the adoption, main challenges and opportunities of GenAI and LLMs in business, the search terms used were:

(“genAI” OR “generative artificial intelligence” OR “large language model” OR “large-language model” or LLMs OR chatGPT) AND (“business” OR “entrepreneurship” OR “entrepreneur” OR “company” OR “industry” OR “enterprise” OR opportunit* OR concern)

In the case of publications, the search query included title and abstract. The time span for data collection extends from the origin of times to October 29th for patents, and to November 10th 2023 for academic publications.

As part of the arranging phase, 792 articles remained in the scope of the research, after removing duplicated articles, thesis, opinion articles and those whose DOI (Digital Object Identifier) was not valid. The preprocessing of texts included the removal of stop words, which are words that do not add any meaning to sentences (e.g. prepositions, pronouns, etc.), emojis and URLs, and the harmonisation of words through lemmatisation.

3.2. Co-occurrence graph-based clustering

A text co-occurrence network serves as a representation of relationships between words or terms, derived from their co-occurrence in a given body of text. In this network, nodes denote words or terms, and edges signify instances where these words appear together in proximity within a document or set of documents. Applying algorithms for community detection in nodes reveals main clusters and the relevant terms within each cluster. Analysis of terms within these clusters can yield valuable insights into the thematic structure of the text, useful for topic modelling (Qiang et al., 2020).

In this research, concerning publications, a co-occurrence network was generated from the corpus formed by titles and abstracts. This process considered a sliding window of three terms and a weight decreasing from three depending on the distance, in line with Paranyushkin (2011). Following the approach of Cano-Marin et al. (2023), clusters were identified by grouping nodes based on modularity, a quality measure for graph clustering (Brandes et al., 2007). The most relevant nodes were identified using eigenvector centrality, a metric that measures the level of influence or prestige of a node in a network (Garg & Kumar, 2018).

In the case of patents, a network was generated by connecting different categories used in patent classification, interconnecting all categories within a given patent. Analogously, clusters were identified based on the modularity class of nodes, and nodes were sorted by degree to identify the most relevant areas of application.

The tools employed for this analysis were NetworkX, a Python library for complex network generation and analysis, and Gephi, an open-source software for graph and network analysis (Bastian et al., 2009).

4. Results and discussion

The interest in GenAI is experiencing rapid growth. As depicted in Figure 2, the initiation of ChatGPT in November 2022 served as the catalyst for the evolution of this field, igniting an increase in the number of publications and patents since then.

4.1. Literature review

In Table 1, the most cited academic publications on the usage of GenAI in businesses are presented. In response to RQ1, the wide range of applications of this technology across different fields is highlighted, including healthcare, education, academia and computer science, among others. In line with Figure 2, the interest in this matter is exponentially increasing, and numerous articles have garnered a substantial number of citations.

In Table 2, the top-10 articles sorted by normalised citations (C_n) are presented. C_n is a measure that normalises the citations of an article since it was published until the present date with the aim of identifying rising-star publications with potential of becoming highly-cited papers. The most cited article (Sallam, 2023) is included as a reference in the first row. The study conducted by Noy et al. (2023) is emphasised as the one with the highest potential based on this metric. This article summarises its potential impact on the productivity of workers through the utilisation of ChatGPT. Once again, emphasis is placed on the myriad of applications across industries.

In response to RQ2, a co-occurrence network was created from the titles and abstracts of the articles in the scope of the literature review, as described in the Methodology section. The resulting clusters are presented in Table 3. The most dominant cluster is natural language generation and understanding, representing 17.93% of the total number of nodes. This reflects the ability of GenAI to process and effectively generate human-like content. Secondly, GenAI opportunities and risks cluster represents 16.1%, identifying the interest in evaluating this technology in business contexts, aiming at improving productivity, automating business activities and enhancing efficiency, at the same time that risks and challenges are observed and mitigated. Thirdly, the use of GenAI in healthcare (14.34%), in line with the literature review, is outlined. GenAI has the potential to transform how information is generated and consumed. Considering that 80% of clinical data is unstructured (Li et al., 2022), including the information available used in medical research and patient care available in Electronic Health Records (EHRs), we can see the potential of its application in healthcare and medicine. With 12.21% of nodes, the potential of using GenAI across industries is present, especially motivated by the development of chat-based applications. In addition, the wide range of applications/tasks where GenAI can be applied along with the identification of ethical risks and limitations are also captured through this topic modelling.

4.2. Patents

To answer RQ3, a network analysis is completed on the graph which was built from the co-occurrence of the categories of patents in Google Patents. 223 unique categories were identified in 213 patents, out of the 409 initially collected. In Figure 3, the representation of the network is depicted. The different clusters are identified and classified in Table 4.

Overall, Table 4 presents a comprehensive breakdown of patent categories, each contributing to the broader landscape of GenAI applications in various business domains. First of all, the main finding of the patent analysis is that there are ongoing efforts to build the foundations of GenAI systems, as represented in the clusters Machine learning, Advanced Natural Language Processing, Communication and Interface Technologies, Secure Transaction Management and Human-Robot Interaction and Language Processing, aiming at improving the efficiency, security and performance of the existing models while mitigating their risks. The application of GenAI in healthcare is observed, in line with the previous research.

5. Challenges in Generative AI in business

Exploring the adoption of GenAI in organisations unveils a complex landscape. Given this complexity, consequently, Table 5 proposes a categorisation of risks and challenges arising from GenAI adoption in businesses, with each category defined by its distinctive set of complexities and considerations. In response to RQ4, the challenges are categorised into strategic, technical, operational, and ethical and regulatory dimensions. From the specifics of design and implementation to day-to-day operational hurdles, ethical considerations and strategic imperatives, each dimension adds a unique layer to the obstacles to overcome. The strategic dimension addresses challenges related to long-term planning and governance, in line with business needs. The technical dimension focuses on issues in the design and implementation phases of GenAI, while the operational dimension deals with challenges in day-to-day life cycle. Lastly, the fourth dimension, covering ethical and regulatory considerations, involves navigating challenges and risks related to ethical concerns and ensuring compliance with legal and regulatory frameworks throughout GenAI’s development, deployment and operations. In summary, this classification serves as a guide, providing insights into the challenges defining the journey to unlock the full potential of GenAI while ensuring responsible and sustainable deployment.

In the strategic dimension, challenges unfold at the intersection of planning, governance, and long-term considerations. Prioritisation becomes a strategic move, aligning GenAI projects with organisational goals and objectives, given the investment and resources required to implement a project involving digital technologies like AI (Richard et al., 2021). Thus, ensuring a measurable return on investment (ROI) is essential, and a thorough evaluation of the value proposition of GenAI initiatives should be completed before implementation to maximise the benefits of investments (Ustundag et al., 2018). In addition, sustainability considerations urge a balance between technological advancement and environmental responsibility, given the water consumption and carbon footprint associated with computing resources required for the training and usage of LLMs in training and inference processes, and the required infrastructure to run those models (Hacker et al., 2023). The swift pace of technological advancements increases complexity, as GenAI capabilities slowly transition into commodities, which may end up not offering any competitive advantage, in line with Abonamah et al. (2021), and models swiftly become obsolete. To represent this unprecedented pace of change, ChatGPT took only five days to reach 1 million users and surpassed 100 million users in less than two months (Statista, 2022). In addition, the choice between open-source and proprietary models demands strategic deliberation. Furthermore, forming strategic alliances and partnerships becomes a lever for enhancing GenAI capabilities, and the selection of vendors and partners may lead to lock-in situations (Schneckenberg et al., 2021), thus needing a delicate dance to manage dependency risks. At organisational level, strong leadership support and a vision are essential (Smith & Green, 2018), providing the impetus for GenAI initiatives across all levels of the organisation. The demand for agility underscores the need to adapt swiftly to changes in the GenAI landscape and technological advancements. Cultural and organisational readiness introduces a nuanced dimension, underscoring the significance of harmonising organisational principles with the integration of GenAI. This highlights the necessity for implementing upskilling programs to internally cultivate the skills required to operate this technology (Jarrahi et al., 2023). Additionally, awareness campaigns are crucial for identifying new opportunities and ensuring a comprehensive understanding of both the benefits and limitations. However, the backdrop of a scarcity of skills and expertise poses a foreseeable challenge, demanding strategies to bridge the talent gap (Budhwar et al., 2023; Ogunrinde, 2022).

From a technical standpoint, ensuring the scalability of GenAI systems to meet expanding data volumes and user demands is a fundamental concern, aligned with the implementation of other digital transformation technologies (Mielli & Bulanda, 2019). Additionally, a paramount consideration such as prioritising security, necessitates a vigilant stance against cyber threats (such as malware code generation, hacking, intelligence gathering, and phishing attacks) to safeguard the confidentiality and integrity of sensitive data (Okey et al., 2023). In the integration of GenAI into the digital ecosystems of companies, achieving interoperability with existing legacy systems requires a delicate balance, demanding seamless compatibility. At its core, a robust IT infrastructure creates a backbone, needing continuous investment and attention. In the case of LLMs, the imperative of AI explainability compels efforts to enhance transparency in GenAI decision-making processes, promoting better understanding and acceptance. Navigating through these technical challenges also involves addressing the inherent limited predictability and ensuring the quality and reliability of the data that underpins GenAI functionality.

Operational challenges, on the other hand, unfold in the day-to-day implementation and maintenance of GenAI systems. Cost management and FinOps, which are the financial management practices to effectively control cloud costs (Storment & Fuller, 2023), become critical considerations for the efficient stewardship of resources throughout the development and deployment phases of GenAI capabilities, given the high cost of training or inference for proprietary models (Barreto et al., 2023) such as LLMs API calls cost (Chen et al., 2023). The call for modularity and flexibility echoes in the design ethos, advocating for systems that can adapt to changes and updates with ease (Dano, 2019). The absence of benchmarks underscores the need for a standardised yardstick to assess the performance and effectiveness of GenAI systems. The governance of data, marked by effective frameworks, emerges as a key component, along with the promotion of reusability and the sharing of best practices (Janssen et al., 2020). Maintenance is a persistent challenge, requiring measures to ensure the ongoing functionality and performance of deployed GenAI systems.

Ethical and regulatory challenges extend beyond technical and operational concerns. The imperative to mitigate bias and ensure fairness in GenAI systems is a moral compass that guides development efforts (Weidinger et al., 2021), highlighting the need to mitigate biases in training data and to provide transparency in decision-making algorithms (Oduoye et al., 2023). AI explainability has become a cornerstone, addressing concerns around the opacity of decision-making processes (Dwivedi et al., 2023a). This lack of model transparency can generate instances of hallucinations, where systems generate inaccurate or misleading outputs, necessitating careful scrutiny and correction (Zhao et al., 2023a). In this line, toxicity should also be prevented, in the form of harmful behaviour, and it underscores the importance of fostering responsible AI (RAI) practices and protecting intellectual property associated with GenAI innovations, in line with the ethical guidelines defined by the European Commission for trustworthy AI (Hleg, 2019). In addition, the need for understanding black-box models and getting clarity on how the models were trained, which data was used in the training and how data collected in the interactions with the models is used for training future models, is essential. Finally, compliance with legal and regulatory requirements is paramount, encompassing considerations such as copyright, protection of personal identifiable information (PII) in line with existing regulations (e.g. GDPR, HIPAA), and accountability mechanisms, aligned with the research of Lucchi (2023). Lastly, it should be considered that ongoing regulations are being developed to regulate the usage of GenAI and LLMs (Hacker et al., 2023).

In summary, the success of adopting GenAI requires an interdisciplinary approach, fostering collaboration and strategic alignment across these multifaceted dimensions (strategic, technical, operational and ethical). Only through such comprehensive engagement can the full potential of GenAI be unlocked, ensuring responsible, sustainable, and impactful deployment.

6. Conclusions

This research provides a perspective on implementing and developing GenAI capabilities within digital systems that support business processes. The review of the state of the art incorporates scientific literature and patents, offering insights into innovation from both academia and industry. In general, AI, particularly GenAI, are advancing rapidly with applications spanning various industries, especially in software engineering, healthcare, business management and education/academia.

Furthermore, the article introduces a novel prescriptive classification of challenges and risks arising from the implementation and use of these technologies. This classification aims to assist business and IT leaders in making effective decisions from strategic, technical, operational, and ethical standpoints while addressing existing challenges.

6.1. Limitations

This study has certain limitations that need to be acknowledged. Firstly, the choice of data sources, namely Web of Science and Google Patents, along with the study period extending until November 2023, might have resulted in the exclusion of articles and patents that were recently published, particularly considering the swift evolution of GenAI. Another limitation arises from the research’s focus on the English language, potentially introducing bias and excluding pertinent information available in other languages. It should be considered that the number of citations used in this research was updated in November 2023.

Declarations

The authors have no competing interests to declare that are relevant to the content of this article. This research received no external funding.

References

Abonamah, A. A., Tariq, M. U., & Shilbayeh, S. (2021). On the Commoditization of Artificial Intelligence. Frontiers in Psychology, 12 696346 https://doi.org/10.3389/fpsyg.2021.696346

Ahamat, A., & Sin, G. K. W. (2022). Developing a business start-up model for technopreneurs. International Journal of Technoentrepreneurship, 4(3), 198-218. https://doi.org/10.1504/IJTE.2022.127155

Ahmed, I., Jeon, G., & Piccialli, F. (2022). From artificial intelligence to explainable artificial intelligence in industry 4.0: a survey on what, how, and where. IEEE Transactions on Industrial Informatics, 18(8), 5031-5042. https://doi.org/10.3145/epi.2014.nov.04

Akter, S., Hossain, M. A., Sajib, S., Sultana, S., Rahman, M., Vrontis, D., & McCarthy, G. (2023). A framework for AI-powered service innovation capability: Review and agenda for future research. Technovation, 125 102768 https://doi.org/10.1016/j.technovation.2023.102768

Barreto, F., Moharkar, L., Shirodkar, M., Sarode, V., Gonsalves, S., & Johns, A. (2023, February). Generative Artificial Intelligence: Opportunities and Challenges of Large Language Models. International Conference on Intelligent Computing and Networking (pp. 545-553). Singapore: Springer Nature Singapore. https://doi.org/10.1007/978-981-99-3177-4_41

Bastian, M., Heymann, S., & Jacomy, M. (2009, March). Gephi: an open source software for exploring and manipulating networks. Proceedings of the International AAAI Conference on Web and Social Media (Vol. 3, No. 1, pp. 361-362). https://doi.org/10.1609/icwsm.v3i1.13937

Brandes, U., Delling, D., Gaertler, M., Gorke, R., Hoefer, M., Nikoloski, Z., & Wagner, D. (2007). On modularity clustering. IEEE Transactions on Knowledge and Data Engineering, 20(2), 172-188. https://doi.org/10.1109/TKDE.2007.190689

Breuer, T., Schaer, P., & Tunger, D. (2022). Relevance assessments, bibliometrics, and altmetrics: a quantitative study on PubMed and arXiv. Scientometrics, 127(5), 2455-2478. https://doi.org/10.1007/s11192-022-04319-4

Budhwar, P., Chowdhury, S., Wood, G., Aguinis, H., Bamber, G. J., Beltran, J. R., ... & Varma, A. (2023). Human resource management in the age of generative artificial intelligence: Perspectives and research directions on ChatGPT. Human Resource Management Journal, 33(3), 606-659. https://doi.org/10.1111/1748-8583.12524

Cano-Marin, E., Mora-Cantallops, M., & Sanchez-Alonso, S. (2023). Prescriptive graph analytics on the digital transformation in healthcare through user-generated content. Annals of Operations Research, 1-25. https://doi.org/10.1007/s10479-023-05495-z

Cascella, M., Montomoli, J., Bellini, V., & Bignami, E. (2023). Evaluating the feasibility of ChatGPT in healthcare: an analysis of multiple clinical and research scenarios. Journal of Medical Systems, 47(1)33 https://doi.org/10.1007/s10916-023-01925-4

Chen, L., Zaharia, M., & Zou, J. (2023). FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performance. arXiv preprint arXiv:2305.05176. https://doi.org/10.48550/arXiv.2305.05176

Cotton, D. R., Cotton, P. A., & Shipway, J. R. (2023). Chatting and cheating: Ensuring academic integrity in the era of ChatGPT. Innovations in Education and Teaching International, 1-12. https://doi.org/10.1080/14703297.2023.2190148

Dano, E. B. (2019, October). Importance of Reuse and Modularity in System Architecture. 2019 International Symposium on Systems Engineering (ISSE) (pp. 1-8). IEEE. https://doi.org/10.1109/ISSE46696.2019.8984472

Daun, M., & Brings, J. (2023, June). How ChatGPT will change software engineering education. Proceedings of the 2023 Conference on Innovation and Technology in Computer Science Education V. 1 (pp. 110-116). https://doi.org/10.1145/3587102.3588815

Dave, T., Athaluri, S. A., & Singh, S. (2023). ChatGPT in medicine: an overview of its applications, advantages, limitations, future prospects, and ethical considerations. Frontiers in Artificial Intelligence, 6 1169595 https://doi.org/10.3389/frai.2023.1169595

Du, H., Teng, S., Chen, H., Ma, J., Wang, X., Gou, C., ... & Wang, F. Y. (2023). Chat with ChatGPT on intelligent vehicles: An IEE TIV perspective. IEEE Transactions on Intelligent Vehicles. https://doi.org/10.1109/TIV.2023.3253281

Dwivedi, R., Dave, D., Naik, H., Singhal, S., Omer, R., Patel, P., ... & Ranjan, R. (2023a). Explainable AI (XAI): Core ideas, techniques, and solutions. ACM Computing Surveys, 55(9), 1-33. https://doi.org/10.1145/3561048

Dwivedi, Y. K., Kshetri, N., Hughes, L., Slade, E. L., Jeyaraj, A., Kar, A. K., ... & Wright, R. (2023b). “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

Elia, G., Raguseo, E., Solazzo, G., & Pigni, F. (2022). Strategic business value from big data analytics: An empirical analysis of the mediating effects of value creation mechanisms. Information & Management, 59(8)103701 https://doi.org/10.1016/j.im.2022.103701

Enholm, I. M., Papagiannidis, E., Mikalef, P., & Krogstie, J. (2022). Artificial intelligence and business value: A literature review. Information Systems Frontiers, 24(5), 1709-1734. https://doi.org/10.1007/s10796-021-10186-w

Eysenbach, G. (2023). The role of ChatGPT, generative language models, and artificial intelligence in medical education: a conversation with ChatGPT and a call for papers. JMIR Medical Education, 9(1), e46885. https://doi.org/10.2196/46885

Faccia, A., Cavaliere, L. P. L., Petratos, P., & Mosteanu, N. R. (2022, August). Unstructured Over Structured, Big Data Analytics and Applications In Accounting and Management. Proceedings of the 2022 6th International Conference on Cloud and Big Data Computing (pp. 37-41). https://doi.org/10.1145/3555962.3555969

Grünebaum, A., Chervenak, J., Pollet, S. L., Katz, A., & Chervenak, F. A. (2023). The exciting potential for ChatGPT in obstetrics and gynecology. American Journal of Obstetrics and Gynecology, 228(6), 696-705. https://doi.org/10.1016/j.ajog.2023.03.009

Guo, J., Cui, L., Sun, S. L., & Zou, B. (2022). How to innovate continuously? Conceptualizing generative capability. Journal of Innovation & Knowledge, 7(2)100177 https://doi.org/10.1016/j.jik.2022.100177

Hacker, P. (2023). Sustainable AI regulation. arXiv preprint arXiv:2306.00292. https://doi.org/10.48550/arXiv.2306.00292

Hacker, P., Engel, A., & Mauer, M. (2023, June). Regulating ChatGPT and other large generative AI models. Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Transparency (pp. 1112-1123). https://doi.org/10.1145/3593013.3594067

Handa, P., Chhabra, D., Goel, N., & Krishnan, S. (2023). Exploring the role of ChatGPT in medical image analysis. Biomedical Signal Processing and Control, 86 105292 https://doi.org/10.1016/j.bspc.2023.105292

Hassani, H., & Silva, E. S. (2023). The role of ChatGPT in data science: how ai-assisted conversational interfaces are revolutionizing the field. Big Data and Cognitive Computing, 7(2), 62. https://doi.org/10.3390/bdcc7020062

Heredia, J., Castillo-Vergara, M., Geldes, C., Gamarra, F. M. C., Flores, A., & Heredia, W. (2022). How do digital capabilities affect firm performance? The mediating role of technological capabilities in the “new normal”. Journal of Innovation & Knowledge, 7(2) 100171 https://doi.org/10.1016/j.jik.2022.100171

Hleg, A. I. (2019). Ethics guidelines for trustworthy AI. B-1049 Brussels. Access date: 10 Dec 2023 https://digital-strategy.ec.europa.eu/en/library/ethics-guidelines-trustworthy-ai

Garg, M., & Kumar, M. (2018). The structure of word co-occurrence network for microblogs. Physica A: Statistical Mechanics and its Applications, 512, 698-720. https://doi.org/10.1016/j.physa.2018.08.002

Jalil, S., Rafi, S., LaToza, T. D., Moran, K., & Lam, W. (2023, April). ChatGPT and software testing education: Promises & perils. 2023 IEEE international conference on software testing, verification and validation workshops (ICSTW) (pp. 4130-4137). https://doi.org/10.1109/ICSTW58534.2023.00078

Janssen, M., Brous, P., Estevez, E., Barbosa, L. S., & Janowski, T. (2020). Data governance: Organizing data for trustworthy Artificial Intelligence. Government Information Quarterly, 37(3) 101493 https://doi.org/10.1016/j.giq.2020.101493

Jarrahi, M. H., Askay, D., Eshraghi, A., & Smith, P. (2023). Artificial intelligence and knowledge management: A partnership between human and AI. Business Horizons, 66(1), 87-99. https://doi.org/10.1016/j.bushor.2022.03.002

Kanbach, D. K., Heiduk, L., Blueher, G., Schreiter, M., & Lahmann, A. (2023). The GenAI is out of the bottle: generative artificial intelligence from a business model innovation perspective. Review of Managerial Science, 1-32. https://doi.org/10.1007/s11846-023-00696-z

Kasneci, E., Seßler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., ... & Kasneci, G. (2023). ChatGPT for good? On opportunities and challenges of large language models for education. Learning and individual differences, 103 102274 https://doi.org/10.1016/j.lindif.2023.102274

Li, I., Pan, J., Goldwasser, J., Verma, N., Wong, W. P., Nuzumlalı, M. Y., ... & Radev, D. (2022). Neural natural language processing for unstructured data in electronic health records: A review. Computer Science Review, 46 100511 https://doi.org/10.1016/j.cosrev.2022.100511

Liu, Y., Han, T., Ma, S., Zhang, J., Yang, Y., Tian, J., ... & Ge, B. (2023). Summary of chatgpt-related research and perspective towards the future of large language models. Meta-Radiology 100017 https://doi.org/10.1016/j.metrad.2023.100017

Lucchi, N. (2023). ChatGPT: A Case Study on Copyright Challenges for Generative Artificial Intelligence Systems. European Journal of Risk Regulation, 1-23. https://doi.org/10.1017/err.2023.59

Madani, A., Krause, B., Greene, E. R., Subramanian, S., Mohr, B. P., Holton, J. M., ... & Naik, N. (2023). Large language models generate functional protein sequences across diverse families. Nature Biotechnology, 1-8. https://doi.org/10.1038/s41587-022-01618-2

Májovský, M., Černý, M., Kasal, M., Komarc, M., & Netuka, D. (2023). Artificial Intelligence Can Generate Fraudulent but Authentic-Looking Scientific Medical Articles: Pandora’s Box Has Been Opened. Journal of Medical Internet Research, 25, e46924. https://doi.org/10.2196/46924

Mariani, M., & Dwivedi, Y. K. (2024). Generative artificial intelligence in innovation management: A preview of future research developments. Journal of Business Research, 175 114542 https://doi.org/10.1016/j.jbusres.2024.114542

Martín-Martín, A., Thelwall, M., Orduna-Malea, E., & Delgado López-Cózar, E. (2021). Google Scholar, Microsoft Academic, Scopus, Dimensions, Web of Science, and OpenCitations’ COCI: a multidisciplinary comparison of coverage via citations. Scientometrics, 126(1), 871-906. https://doi.org/10.1007/s11192-020-03690-4

Martínez, J. M. G., Carracedo, P., Comas, D. G., & Siemens, C. H. (2022). An analysis of the blockchain and COVID-19 research landscape using a bibliometric study. Sustainable Technology and Entrepreneurship, 1(1)100006 https://doi.org/10.1016%2Fj.stae.2022.100006

Meskó, B., & Topol, E. J. (2023). The imperative for regulatory oversight of large language models (or generative AI) in healthcare. NPJ Digital Medicine, 6(1) 120 https://doi.org/10.1038/s41746-023-00873-0

Mielli, F., & Bulanda, N. (2019, April). Digital transformation: Why projects fail, potential best practices and successful initiatives. 2019 IEEE-IAS/PCA Cement Industry Conference (IAS/PCA) (pp. 1-6). IEEE. https://doi.org/10.1109/CITCON.2019.8729105

Möhring, M., Keller, B., Schmidt, R., Schönitz, F., Mohr, F., & Scheuerle, M. (2022, September). Analytics in industry 4.0: investigating the challenges of unstructured data. International Conference on Business Informatics Research (pp. 113-125). Cham: Springer International Publishing. https://doi.org/10.1007/978-3-031-16947-2_8

Noruzi, A., & Abdekhoda, M. (2014). Google Patents: The global patent search engine. Webology, 11(1).

Noy, S., & Zhang, W. (2023). Experimental evidence on the productivity effects of generative artificial intelligence. Available at SSRN, 4375283. https://doi.org/10.2139/ssrn.4375283

Oduoye, M. O., Javed, B., Gupta, N., & Sih, C. M. V. (2023). Algorithmic Bias and Research integrity; the role of non-human authors in shaping scientific knowledge with respect to Artificial Intelligence (AI); a perspective. International Journal of Surgery (London, England). https://doi.org/10.1097/js9.0000000000000552

Oesterreich, T. D., Anton, E., & Teuteberg, F. (2022). What translates big data into business value? A meta-analysis of the impacts of business analytics on firm performance. Information & Management, 59(6)103685 https://doi.org/10.3390/info14020085

Ogunrinde, A. (2022). The Effectiveness of Soft Skills in Generating Dynamic Capabilities in ICT companies. ESIC Market, 53(3), e286-e286. https://doi.org/10.7200/esicm.53.286

Okey, O. D., Udo, E. U., Rosa, R. L., Rodríguez, D. Z., & Kleinschmidt, J. H. (2023). Investigating ChatGPT and cybersecurity: A perspective on topic modeling and sentiment analysis. Computers & Security, 135 103476 https://doi.org/10.1016/j.cose.2023.103476

OpenAI (2022). ChatGPT. OpenAI. 2022. Access date: 10 Dec 2023. https://openai.com/blog/chatgpt

Paranyushkin, D. (2011). Identifying the pathways for meaning circulation using text network analysis. Nodus Labs, 26, 1-26.

Paul, J., Lim, W. M., O’Cass, A., Hao, A. W., & Bresciani, S. (2021). Scientific procedures and rationales for systematic literature reviews (SPAR-4-SLR). International Journal of Consumer Studies, 45(4), O1-O16. https://doi.org/10.1111/ijcs.12695

Peres, R., Schreier, M., Schweidel, D., & Sorescu, A. (2023). On ChatGPT and beyond: How generative artificial intelligence may affect research, teaching, and practice. International Journal of Research in Marketing, 40(2), 269-275. https://doi.org/10.1016/j.ijresmar.2023.03.001

Perifanis, N. A., & Kitsios, F. (2023). Investigating the influence of artificial intelligence on business value in the digital era of strategy: A literature review. Information, 14(2) 85 https://doi.org/10.3390/info14020085

Perkins, M. (2023). Academic Integrity considerations of AI Large Language Models in the post-pandemic era: ChatGPT and beyond. Journal of University Teaching & Learning Practice, 20(2) 07 https://doi.org/10.53761/1.20.02.07

Qiang, J., Qian, Z., Li, Y., Yuan, Y., & Wu, X. (2020). Short text topic modeling techniques, applications, and performance: a survey. IEEE Transactions on Knowledge and Data Engineering, 34(3), 1427-1445. https://doi.org/10.1109/TKDE.2020.2992485

Ray, P. P. (2023). ChatGPT: A comprehensive review on background, applications, key challenges, bias, ethics, limitations and future scope. Internet of Things and Cyber-Physical Systems. https://doi.org/10.1016/j.iotcps.2023.04.003

Richard, S., Pellerin, R., Bellemare, J., & Perrier, N. (2021). A business process and portfolio management approach for Industry 4.0 transformation. Business Process Management Journal, 27(2), 505-528. https://doi.org/10.1108/BPMJ-05-2020-0216

Salah, M., Al Halbusi, H., & Abdelfattah, F. (2023). May the force of text data analysis be with you: Unleashing the power of generative AI for social psychology research. Computers in Human Behavior: Artificial Humans 100006 https://doi.org/10.1016/j.chbah.2023.100006

Sallam, M. (2023, March). ChatGPT utility in healthcare education, research, and practice: systematic review on the promising perspectives and valid concerns. Healthcare (Vol. 11, No. 6, p. 887). MDPI. https://doi.org/10.3390/healthcare11060887

Schneckenberg, D., Benitez, J., Klos, C., Velamuri, V. K., & Spieth, P. (2021). Value creation and appropriation of software vendors: A digital innovation model for cloud computing. Information & Management, 58(4)103463 https://doi.org/10.1016/j.im.2021.103463

Smith, A. M., & Green, M. (2018). Artificial intelligence and the role of leadership. Journal of Leadership Studies, 12(3), 85-87. https://doi.org/10.1002/jls.21605

Statista (2022, December). Adoption rate for major milestone internet-of-things services and technology in 2022, in days. Statista. Access Date: 15 Nov 2023. Retrieved from https://www.statista.com/statistics/1360613/adoption-rate-of-major-iot-tech/

Storment, J. R., & Fuller, M. (2023). Cloud FinOps. O’Reilly Media, Inc.

Tagscherer, F., & Carbon, C. C. (2023). Leadership for successful digitalization: A literature review on companies’ internal and external aspects of digitalization. Sustainable Technology and Entrepreneurship100039 https://doi.org/10.1016/j.stae.2023.100039

Ustundag, A., Cevikcan, E., Isikli, E., Yanik, S., Cevikcan, E., & Ustundag, A. (2018). Project portfolio selection for the digital transformation era. Industry 4.0: Managing the digital transformation, 105-121. https://doi.org/10.1007/978-3-319-57870-5_6

Vaishya, R., Misra, A., & Vaish, A. (2023). ChatGPT: Is this version good for healthcare and research? Diabetes & Metabolic Syndrome: Clinical Research & Reviews, 17(4) 102744 https://doi.org/10.1016/j.dsx.2023.102744

Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł. & Polosukhin, I. (2017). Attention is all you need. Advances in neural information processing systems30 https://doi.org/10.48550/arXiv.1706.03762

Wang, F. Y., Yang, J., Wang, X., Li, J., & Han, Q. L. (2023a). Chat with ChatGPT on industry 5.0: Learning and decision-making for intelligent industries. IEEE/CAA Journal of Automatica Sinica, 10(4), 831-834. https://doi.org/10.1109/JAS.2023.123552

Wang, S., Sun, X., Li, X., Ouyang, R., Wu, F., Zhang, T., ... & Wang, G. (2023b). GPT-ner: Named entity recognition via large language models. arXiv preprint arXiv:2304.10428. https://doi.org/10.48550/arXiv.2304.10428

Weidinger, L., Mellor, J., Rauh, M., Griffin, C., Uesato, J., Huang, P. S., ... & Gabriel, I. (2021). Ethical and social risks of harm from language models. arXiv preprint arXiv:2112.04359. https://doi.org/10.48550/arXiv.2112.04359

Zhang, Y., Zhang, M., Li, J., Liu, G., Yang, M. M., & Liu, S. (2021). A bibliometric review of a decade of research: Big data in business research–Setting a research agenda. Journal of Business Research, 131, 374-390. https://doi.org/10.1016/j.jbusres.2020.11.004

Zhao, A., & Wu, Y. (2023). Future implications of ChatGPT in pharmaceutical industry: drug discovery and development. Frontiers in Pharmacology, 14 1194216 https://doi.org/10.3389/fphar.2023.1194216

Zhao, H., Chen, H., Yang, F., Liu, N., Deng, H., Cai, H., ... & Du, M. (2023a). Explainability for large language models: A survey. arXiv preprint arXiv:2309.01029. https://doi.org/10.48550/arXiv.2309.01029

Zhao, R., Chen, H., Wang, W., Jiao, F., Do, X. L., Qin, C., ... & Joty, S. (2023b). Retrieving multimodal information for augmented generation: A survey. arXiv preprint arXiv:2303.10868. https://doi.org/10.48550/arXiv.2303.10868

Zhou, C., Li, Q., Li, C., Yu, J., Liu, Y., Wang, G., ... & Sun, L. (2023). A comprehensive survey on pretrained foundation models: A history from BERT to ChatGPT. arXiv preprint arXiv:2302.09419. https://doi.org/10.48550/arXiv.2302.09419

_______________________________

E-mail: enrique.canom@uah.es

ISSN 0212-1867 / e-ISSN 1989-3574

CC-BY

ESIC University

https://revistasinvestigacion.esic.edu/esicmarket

How to cite:

Cano-Marin, E. (2024). The transformative potential of Generative Artificial Intelligence (GenAI) in business: a text mining analysis on innovation data sources. ESIC Market. Economics and Business Journal, 55(2), e333. DOI: 10.7200/esicm.55.333