Ghostwriters in the machine: Openly appreciating AI tools and humans who helped us

摘要

Last year I wrote the JOB ARCDI editorial about my excitement over the artificial intelligence (AI) tool ChatGPT (Dasborough, 2023). To my great surprise, some scholars read the editorial, incorporating and citing it in their own work! Back when that editorial was first published, ChatGPT was relatively new in our field and not many scholars in OB were using it. However, today, it is a very different story. Many OB scholars are using ChatGPT regularly to assist them with fulfilling their research, teaching, and service commitments. The popularity of ChatGPT has grown extremely quickly (faster than Instagram and TikToc!; Berg et al., 2023), and there has even been a ChatGPT review article published (Ayinde et al., 2023). ChatGPT is not the only AI tool that scholars have available to use. Today, a plethora of AI tools now help make scholars live a little easier. As I reviewed the new submissions to the JOB ARCDI over the last few weeks, something that stood out to me was how many of the papers mentioned their use of AI. I had not heard of many of these AI tools before, and I had to look them up online to learn about their capabilities. While some of these AI tools are newly developed, others have been around for a while and have been used frequently in other fields of study (e.g., the tool “Rayyan”; see Ouzzani et al., 2016). It may be that AI tools provide unique forms of assistance for different parts of the review process. For example, DistillerSR primarily helps with screening articles and extracting data from them. EPPI-Reviewer helps managing references, screening articles, coding, and research synthesis. Rayyan, JBI SUMARI, and Covidence suggest which articles to include in a review (data screening) and help produce the review, as well as facilitate cooperation with multiple co-authors conducting reviews. 1 Some AI tools, such as ChatGPT, can be used to brainstorm ideas for papers, create paper outlines, summarize, and paraphrase text, and provide feedback on drafts. Reporting this type of writing assistance may be more challenging than reporting assistance with a specific task such as screening articles; however, it can be done (see Appendix A for an illustration). The second reason that I am pleased to see the reporting of AI tools in JOB ARCDI manuscripts is because it facilitates learning. I love to learn about new tools available, and I feel joy in knowing that this information is also being shared with our JOB audience (see Gagné et al., 2019). If some scholars are benefitting from AI tools, many of which are freely available, then it is leveling the playing field if these tools are then shared with others. When scholars report their selected “tools of the trade,” it helps advance our science and we all benefit from it as scholars. 2 There are eight articles in this issue of ARCDI, each of which may or may not have benefitted from the use of AI tools. As usual, these articles cover a wide range of topics that will hopefully keep the JOB audience engaged and motivated to start new research projects of their own. The first three articles in this issue build on the theme of constant change in the workplace. They provide insights into the intricate relationships between technological advancements, new forms of employment, and the creative processes that drive organizational innovation. First, Bankins and colleagues present a multilevel review of artificial intelligence in organizations, which sheds light on the transformative impact of AI on organizational structures, processes, and human-AI dynamics. This review not only delves into the implications for future research on AI at work but also provides managers with guidance on prioritizing employee well-being while embracing the use of AI. The emergence of AI platforms and algorithms has enabled the automation of certain tasks, leading to the creation of new gig work opportunities, as well as streamlining the matching process of gig workers and employers. 3 The second