Background

Artificial intelligence (AI) has advanced tremendously and revolutionized several industries. AI driven applications has the functional capabilities to generate new content, graphics, audio, and video, is called generative artificial intelligence (AI) [1]. Generative AI are one recent applications of AI which learn from large datasets (training datasets) and then produce unique and innovative content human reply style for the users in response to their queries (also called prompts) [2]. Generative Pre-trained Transformer (ChatGPT) is one example of a generative AI model that has been trained on vast quantities of data to generate textual responses in a chat-like format in response to user inquiry. The continuous research by OpenAI is being conducted to continuously innovate ChatGPT, for instance launch of new multimodal large language model called GPT-4 compared to previous GPT-3.5. The latest model is reported to have better capabilities to provide more accurate responses, better problem solving, and better features, for instance images analysis rather just text analysis (https://openai.com/gpt-4). Innovations in generative AI technology are expected to unfold rapidly, intensifying competition among industry players to develop better generative AI technologies. The market is poised to witness a continuous influx of innovative Generative AI technologies, leading to the emergence of sophisticated applications that either incorporate inbuilt generative AI or are purposefully built using such technologies. These innovations will create great value for its users, for instance, entrepreneurs, who will have the opportunity to navigate the rapidly evolving landscape of such technologies to harness its potential benefits for their ventures. Given that AI is still in its early stages of development, they must also take these challenges into account. The integration of these technologies with business practices seems not to be a future of work problem which is too distant; it is going to happen soon; at least looking at the pace of innovations in Generative AI technologies.

Numerous businesses had been actively using ChatGPT, and it was stated that their use had a positive relationship with the productivity [3], and success of the company [4]. The ChatGPT quality, Information quality, and Service quality is reported to have positive impact on user benefits and their satisfaction. This ultimately positively impact the organizational performance [3]. For instance, the system quality, information quality and service quality factors can result in accurate and timely market knowledge for the marketing team, thereby enhancing the user benefits their satisfaction levels. This may result in the better marketing decisions, for instance, better pricing decisions, and better impact, for instance, higher sales and business performance. Additionally, the positive impacts of adoption of ChatGPT includes cost reduction and operational efficiency [5, 6] support for Business Model Innovations [7], and possibility of integration in one of the crucial areas of Research and Innovation [8]. Although this technology is still in its early phases, it appears to be worthwhile for businesses, including startups to investigate further. Startups employ a range of effortful methods to perform market research, expert and customer interviews, internet research tools, and analysis of secondary materials, for instance, published market reports. For startups with limited resources, ChatGPT may be a useful knowledge collaborator. It can assist the startup team in utilizing the information generated by this technology to augment the data obtained from their current market research procedures. More thorough data collecting, and analysis could help improve the multiple points of view that were produced, which would result in rational business model innovations.

The literature has examined the adoption of AI technology by either extending theoretical models, such as the Unified Theory of Acceptance and Use of Technology (UTAT) model, and Technology Acceptance Model (TAM) [9], or by proposing new models, such as the Task-oriented AI Acceptance (T-AIA) model [10] and Artificially Intelligent Device Use Acceptance model (AIDUA) [11]. Nonetheless, research focused solely on the acceptance of generative AI technologies—such as ChatGPT—has examined the acceptance from the perspectives of Students [12–16], knowledge workers [17], healthcare professionals [18], and customers [19]. To undertake such investigations, the researchers has extended TAM model [12–15], UTAT Model [16, 17], and AIDUA Model [19] as the research frameworks. The dearth of prior research on ChatGPT adoption in the startup context, especially from Entrepreneur perspective, highlights the urgent need for a thorough investigation to identify the variables influencing this technological adoption. Determining the factors that contribute to a startup’s successful integration of ChatGPT is critical because they operate under unique working contexts and constraints. This study investigates these characteristics to give entrepreneurs critical insights that will help them make rational decisions, reduce technology related risks having a business impact, and foster their continuous innovations in their entrepreneurial environment. With time, more advanced and innovative Generative AI technologies and applications based on them will hit the market as the AI field develops. The continuous adoption of the emerging Generative AI technologies will help startups gain more and more experience with adoptions, helping them to leverage continuously evolving technological innovation landscape.

Rationale

It is crucial to comprehend the factors that affect startups’ adoption of ChatGPT technology. This will enable the entrepreneurs to make decisions on whether they should adopt the technology. For instance, the results will require startups to consider if their working environment and available resources match the requirements for a successful adoption. Additionally, the results will allow them to decide whether to execute the small-scale, incremental experiments that should be the foundation for adopting AI technology [20, 21]. For example, the entrepreneurs will be able to design small experiments based on the adoption factors and accurately identify the performance criteria for these experimentations. The results have implications for different stakeholders, for instance, Entrepreneurs, Policymakers, technology providers, researchers, and Institutions offering support for entrepreneurs like Academia, Incubators and Accelerators, University libraries, public libraries, chambers of commerce, and foreign embassies. These stakeholders will find the research findings useful in improving their understanding of the adoption determinants and obstacles, which will improve their ability to make decisions about their service innovations and, ultimately improve the environment for entrepreneurship.

Objectives

The primary objective of this study is to ascertain the factors that impact the uptake of ChatGPT technology by startups, anticipate their influence on the triumph of companies, and offer pragmatic suggestions for various stakeholders, including entrepreneurs, and policymakers.

National and local approvals/ethical considerations and approval

The research study is approved by the Institutional Review Board of Winning Scientific Management, Lisbon, Portugal under protocol number 0823J. The research project is also approved by the University of Leicester Ethics Sub-Committee of Science & Engineering, Arts, Humanities and Law (research study design ID: 43400). The research project has also obtained clearance under the Academic Technology Approval Scheme (ATAS) of the UK Government’s Foreign, Commonwealth, and Development Office (FCDO) dated July 20th, 2023.

Data analysis

Partial Least Squares Structural Equation Modeling (PLS-SEM) will be used to analyze the gathered data. PLS-SEM offers a powerful statistical analytic tool to ascertain the relevance and influence of each factor on startups’ adoption of ChatGPT technology, as well as the ability to examine complicated interactions among variables. The study will analyze the structural connections among the suggested constructs, appraise the importance of these connections, and ascertain the degree of their influence. PLS-SEM is selected for the data analysis because this technique is reported to work well for small sample sizes (Entrepreneurs using ChatGPT for business operations are increasing but still comprises small population size), complex models (adoption model is complex with diverse factors and interactions) and does not make any assumptions about data distribution (not sure if Likert scale data collected in the research will be normal or otherwise) [24, 25]. Further, this technique is well suited if the objective of the research is to develop the theory, explain how much of the variability in your dependent variables (constructs) can be accounted for by the independent variables (factors or predictors) in your model, and predicting the relationships between different constructs within the adoption theoretical model [25].

Follow-up

A procedure for follow-up will be instituted to guarantee that the experiences and concerns of the participants are considered. During the Stage two of the research study, during the filling of the survey questionnaire, participants will be given the opportunity to indicate their willingness to engage in follow-up interviews, for instance, exploring reasons behind analysed statistical correlations in research framework or other inquiries, for instance, seeking their general opinion about future implications, by providing their contact information in the questionnaire itself. The goal of these optional follow-up conversations is to gather rich perspectives about the analysed results, especially unexpected outcome situations (if any). This will enable researchers to offer more in-depth explanations of the elements influencing technology adoption (just like case study explanations) that simply survey-based research would not be able to answer (quantitative based). The follow-ups are totally optional, and it is possible to participate in the research without sharing any personal information. Interviews will be scheduled at the convenience of follow-up participants, who will only be contacted by phone, email, or online meetings if that is their chosen method of communication. Researchers will have a deeper grasp of the multiple viewpoints about the research problem through multiple rounds of follow-ups. This will contribute to extending the research study’s outcomes and enhancing its reliability and validity.

Sampling

Purposive or convenience sampling, a non-probability sample technique, will be employed in the study, which will feature startups from the researcher’s professional network (and extended network) in universities, public libraries, accelerators, incubators, and similar institutions. The researcher intends to focus on startups in their professional network (and extended network) that are either considering adopting ChatGPT technology or have rich experiences or prior experience with it. These startups will be picked because they can offer insightful information on the factors driving technology adoption and because they are in line with the goals of the research (they have experience with ChatGPT technology adoption).

ChatGPT is recently introduced in the market and its integration with the business practices is still under active experimentation by the entrepreneurs. Entrepreneurs maybe using ChatGPT in their business operations at varying degrees, for instance, for easy task of drafting marketing messages to knowledge intensive task of market research. Its use is now less widespread, but it is growing as more peer success stories are made public and entrepreneurs are scaling their experimentations with the technology. The population size of entrepreneurs actively using ChatGPT for business operations seems to be smaller but cannot be estimated quantitatively. Assuming the confidence level of 95% and ±5% of the margin errors, the sample size is computed to be 385.

The startups in researchers’ professional network as well in the extended professional network, will receive an invitation to participate along with information on the goals, nature of the research project, and eligibility criteria for participation (active use or having used Generative AI in their business operations). The research project information, for instance, research objectives, voluntary participation, associated risks and benefits, data confidentiality, specifics about personal data collection (if any), safety precautions (refer to the Ethics and Safety Considerations section), and contact persons, will be shared with potential participants to help them make an informed decision about whether or not to take part in the research. The University of Leicester’s standard participant information sheet (annexure-A attached in S1 Annex) will be used to disseminate research project details. The participant information sheet will provide participants with sufficient project-related information to enable them to make an informed decision about their participation in the project. The Participants will also be shared the General Data Protection Regulation (GDPR) privacy notice which applies if they wish to share their personal data. The University of Leicester’s GDPR Privacy Notice explains how it gathers and utilizes participants’ personal data when they participate in the research project (annexure-B attached in S1 Annex). Participants agreeing to participate will provide their written informed consent using the University of Leicester’s standard Informed consent form (annexure-C attached in S1 Annex). After receiving consent, the researcher will give the participating startups structured questionnaires to complete to gather data. The founders of the startup have to option to explicitly give their permission to be contacted in future for necessary follow-ups (refer to the data collection section) by providing their personal information (refer to data storage section for information about retention period). In case they expressed their consent to participate in the follow-up sessions, new participant information sheet (annexure-D attached in S1 Annex), and informed consent form will be shared (annexure-E attached in S1 Annex). The conditions expressed before the start of the survey, for instance data processing, ethical requirements, GDPR notice, will all remain the same for the follow-up sessions.

Data collection

The startup entrepreneurs will be given a standardized questionnaire (which is built based on outcomes of stage one of the research study) to complete to gather data. The informed consent will be taken before distributing the questionnaire (refer to the Ethics and Safety Considerations section). The questionnaire will be created through Google Forms and distributed among the research study participants. To guarantee precision and effectiveness, the answers will be electronically recorded (Google spreadsheet). The questionnaire will also have the fields where participants could enter their organization details, their contact details (in case interested in follow-ups), background knowledge of AI technology, industry context, specific challenges, and success stories about their firms. The survey form will not record any personal information unless and until the participant explicitly feeds the information for follow-ups. This means that even the e-mail ids won’t be recorded but they will have the option to edit their response after submission (they just need to save their survey link with the option “Edit your response”).

Data storage

An electronic database that is password-protected and safe will house all the gathered data. This database will only be accessible to members of the authorized research team. The EU and UK GDPR laws does not specify the time period for which personal data can be retained. Businesses are free to choose how long to keep personal information, for instance, based on the purpose it was collected. The stored personal data will be deleted 6 months after the end of the research study to allow ample time for any follow-ups and research article drafting. The collected data (5-point Likert scale values) as well as personal data (if shared by entrepreneurs) will not be shared with any third parties, even outside EEA. The 5-point Likert scale research data will also be destroyed beyond final publication (publication of Stage two results).

Data analysis

The Partial Least Squares Structural Equation Modeling (PLS-SEM) method will be used to conduct a thorough analysis of the gathered survey data. This approach will evaluate the correlations between the variables, determine the relevance of the results, and quantify the magnitude of the effects. For this, the SmartPLS 3.2.9 software version will be used.

Data handling procedures

Data management will adhere to ethical and privacy standards. The research team leaders will be the only ones with access to the data, as it won’t be available to other parties. Individual data gathered from the survey will not be shared in any follow-up correspondence with respondents. Additionally, the volunteer researchers will only be able to access the 5-point Likert scale values that have been registered for each question in the questionnaire (removing any personal or corporate-specific information) to guarantee that analytic methods are carried out appropriately (refer to the Quality Assurance section). Only when participants voluntarily provide their personal information, it will be gathered (for follow-ups). Specifically, it is feasible to participate in a research study without disclosing any personal information or information unique to the participating business. All submitted information will only be used in aggregate form; no personal information will be disclosed by the research project. Only with consent from the participants (captured in informed consent form) will the specific quotes that they contributed to the questionnaire or in the follow-ups be made public (but anonymised, no personal information will be revealed).

Informed consent

Informed consent will be obtained from research participants before starting data collection. They will receive complete disclosures about the objectives of the study, the research methods to be employed, the non-disclosure of the results, the privacy of their personal information, and how the research study will use their answers. This will be accomplished by sharing participant information sheet, General Data Protection Regulation (GDPR) privacy notice, and Informed consent form to fill. Before taking part, they will have the chance to ask questions (through email) and give their voluntary consent (through signed informed consents). In any event, participants are free to leave the research project at any moment and without giving a reason.

Privacy and confidentiality

To preserve participant privacy, personal information will only be collected (for follow-ups) when participants willingly submit it. The researchers will have no ways for requesting the participants for their personal information (as the option will only be available in survey form and it is impossible to reach out back to them as their emails will not be recorded). It is possible to take part in a research study without giving out any private information or information exclusive to the participating company. No personal information will be revealed by the study project (if shared by participants); all provided data (Likert scale values) will only be utilized in aggregate form. No personal information will be disclosed, for instance in publications, without the consent of the participant. The participants’ approval will be obtained before making public the precise quotes they provided at the time of filling the questionnaire. The research data (Likert scale values and personal information) will only be accessible to leaders of the research team while volunteer researchers (if any) will only have access to 5-point Likert scale values (research data excluding personal information). The personal or company-specific information will be eliminated from the dataset shared with the volunteer researchers.

Minimizing harm

Participants can complete the questionnaire without difficulty or stress because it can be completed in a week at their own pace (possibility of completing in increments in a week). There is no risk of physical injury because the entire process can be completed online from the comfort of the participant’s home or place of business.

Data security

The safety and privacy of participant information will be guaranteed by the implementation of data security measures, such as password protection and secure storage. Only the research project leaders will have access to the storage. To guarantee that no unauthorized individual can access the research data, the workstation used to access it will be password protected. In case the research data do not include any personal information then it will be simply stored on the university cloud storage through OneDrive file hosting service. The OneDrive allows the data to be shared with the peer researcher for research collaboration. It is not possible to access the data without an official login and granting access to the research data. In case personal data is also collected (in case participants shared personal data for follow-ups), the University’s Records Manager will be requested to add it to the official register as per the Regulations the University. To prevent sharing of this data on third-party storage, it will only be shared within the researcher’s password-protected computer memory.