Portal de Periódicos da CAPES

Should We Publish Fewer Papers?

2024; American Chemical Society; Volume: 9; Issue: 8 Linguagem: Inglês

10.1021/acsenergylett.4c01991

2380-8195

Song Jin,

CO2 Reduction Techniques and Catalysts

InfoMetricsFiguresRef. ACS Energy LettersVol 9/Issue 8Article This publication is free to access through this site. Learn More CiteCitationCitation and abstractCitation and referencesMore citation options ShareShare onFacebookX (Twitter)WeChatLinkedInRedditEmailJump toExpandCollapse EditorialAugust 9, 2024Should We Publish Fewer Papers?Click to copy article linkArticle link copied!Song Jin*Song JinDepartment of Chemistry, University of Wisconsin−Madison, 1101 University Avenue, Madison, Wisconsin 53706, United States*[email protected]More by Song Jinhttps://orcid.org/0000-0001-8693-7010Open PDFACS Energy LettersCite this: ACS Energy Lett. 2024, 9, 8, 4196–4198Click to copy citationCitation copied!https://pubs.acs.org/doi/10.1021/acsenergylett.4c01991https://doi.org/10.1021/acsenergylett.4c01991Published August 9, 2024 Publication History Received 23 July 2024Accepted 24 July 2024Published online 9 August 2024Published in issue 9 August 2024editorialCopyright © 2024 American Chemical Society. This publication is available under these Terms of Use. Request reuse permissionsThis publication is licensed for personal use by The American Chemical Society. ACS PublicationsCopyright © 2024 American Chemical SocietySubjectswhat are subjectsArticle subjects are automatically applied from the ACS Subject Taxonomy and describe the scientific concepts and themes of the article.CrystalsEnergyPhysical and chemical processesQuality managementRedox reactionsIt is perhaps an understatement to say that all of us in modern society, especially academic researchers, are overwhelmed. There are always more papers to write, more grant applications to submit, more administrative reports to file, more committees to serve on, more conferences to attend, and more manuscript (and grant) reviews to perform. Particularly on the academic papers, the number of peer-reviewed research publications has been growing rapidly by about 8–9% each year. (1,2) Specifically, as one of the fastest growing research areas, renewal energy research has also witnessed a significant growth in the number of publications, as well as the number of energy research journals. It seems that we are on a perpetual treadmill that is getting faster and faster, and what is worse, it seems that we are powerless to change the course. It is clear that everyone is working harder and harder, trying to make an impact with research efforts, but it is not clear that we are making more progress or we are getting better. This phenomenon seems to fit well into the classic economic theory of "involution". (3)Therefore, I venture to ask the following question to the academic research community: Should we all publish fewer papers? By this proposition, I am not advocating that we all "slack off", but rather that we work harder to make sure we publish fewer but higher quality papers with new and significant scientific insights instead of reporting routine or incremental studies in large numbers of papers. This is by no means a new problem for our time─the issue of "salami" papers has been long known, and my editorial on "sandwich papers" also seemed to resonate with many readers. (4) However, the emerging generative and large language models artificial intelligence (AI) tools, such as ChatGPT, are making this problem even more acute. With the aid of such AI tools, seemingly reasonable research papers (both original research papers and reviews) can now be prepared very quickly (and guidelines are being developed on such practice). (5,6) Alarmingly, some research manuscripts must have been reviewed by using such AI tools already. (7,8) Then, some of us are probably relying on AI to summarize research papers so that we can read more papers faster. (9) With all of this going on, it is pointless for us Earthly beings to compete with AI machines to strive for generating more and more research papers. Instead, we must focus on doing what AI tools cannot do well: creative, original, and significant new research works that really answer some scientific questions and solve previously unsolvable problems. As far as I can tell, language AI tools have not (yet) been very good at addressing such challenges.Figure 1Figure 1. Researchers are overwhelmed by the number of research papers being published these days. (Source: iStock.com/aldegonde)High Resolution ImageDownload MS PowerPoint SlideIt is also quite curious why so many new academic research journals keep popping up every week? As a journal editor myself, I still could not keep up with the dizzying pace of announcements for new journals and cannot make sense out of all of these. Of course, I am not disputing the fact that some research fields experience rapid growth and indeed new publication venues distinct from those for the traditional disciplines are needed to accommodate the explosive growth of research papers, but this is clearly not always the case when one examines the names and scopes of many new journals. Especially now when most of these new journals are online only, what are the differences and benefits in setting up more and more small journals on highly specialized and specific topics (some of which might just publish a few dozen papers a year), instead of just archiving them together in one place? If the growth rate of academic publications is even slower than the growth rate of academic journals, the published papers are going to be increasingly archived in more and more fragmented fashion at different journal Web sites. This is not even mentioning the flood of predatory Open Access journals and publishers that have led to academic fraud and the retraction of a massive number of papers. (10)With the proliferation of new journals, a manuscript can bounce between more and more different journals for repeated resubmissions, which means more editors and reviewers will need to evaluate the same research work over and over again as voluntary service. This undoubtedly contributes to the increased workload for all academic researchers but results in very little real benefit for the community as a whole. Every paper written eventually gets published in some journal, so the fewer rounds of submission it has to go through, the less work for the whole community. This issue of journal proliferation is also related to the issue of having more papers─if the number of publications (regardless of how meaningful and valuable they are) keeps growing quickly, publishers see more content to capture and will be motivated to create more journals. Simply put, if there is more demand, there will be more supply. It is quite easy for us to say "no" to manuscript review requests, but would we resist the temptation to publish in yet another new journal, especially if we are invited to do so? Perhaps the research community (we, the practicing scientists) as a whole needs to have some honest and healthy conversations among ourselves about how we would approach this issue.Since we are all rational scientists, one needs to ask why we could not get out of such unproductive cycles. The conclusion must be that we are all motivated by the evaluation and incentive systems we are in. It is always simpler and easier to just use some quantitative metrics to evaluate the research output and researchers' productivity and impact─the number of publications, the impact factors of the journals in which the research works are published, the citations, etc. So long as evaluation processes exist, it is probably impossible to completely avoid some form of "bean-counting". Therefore, maybe we could try to improve the ways we "count the beans", to provide more incentive for publishing fewer (and hopefully more well-thought-out and higher impact) papers. As flawed as it might be, the Hirsch index (h-factor) (11) has become one of the most commonly used metrics to characterize the scientific output of an individual researcher. Here I argue that the ratio of the H-index to the total number of publications for a given researcher reveals more information about what fraction of the research papers have truly made impact (at least in terms of citations). We could use this ratio as a weighting factor for the original h-index to calculate a "weighted h-index" (= h2/total number of publications) that can reflects the difference between someone who has a very large number of publications with a small fraction of them being highly influential versus someone who has a smaller number of publications but with a larger fraction of them being highly influential, and thus incentivize the latter case. Of course, we could further debate or refine how much this fraction should weigh differently by using the square root or other operation on the fraction of h/total number of publications.I must admit that, as a scientist who is considered reasonably "productive", I have published quite a few papers myself and might well be part of the problem here. If we assume that half of the submitted manuscripts get published in a given journal, for each published manuscript, there would be at least 4 peer reviews conducted in general. I am not sure that I have contributed enough peer reviews commensurate with the number of papers I have published. Feeling the "guilt", I have been increasingly asking myself─Do I really need to write that paper? What differences would my paper make? What scientific (or engineering) problems would my paper help to solve, and what new questions could my paper answer? Are these interesting, meaningful, and significant advances? Do I need to write a review manuscript when I do not have something burning to say and reviews on similar topics already exist? Of course, not all of these questions, especially about the significance and potential impact of a given research paper, could be fully answered a priori, but at least I need to engage in some internal debate about them. Furthermore, with the same body of experimental results, could I summarize them in more concise and efficient, but clear and accurate ways to be published in fewer papers so that future readers' time and effort in reading my papers could be more worthwhile (instead of driving them to use AI tools to process my papers)? I am not sure that I have been making progress in achieving any of these goals above, but I must try, because the alternative is not good. I am also not sure if I am making some useful points in this editorial, but I hope to persuade you to join me in this pursuit of trying to publish fewer papers, because I am convinced that otherwise we are on an unsustainable path.Finally, switching back into my hat as an editor for ACS Energy Letters, you might ask: if everyone tries to write fewer papers, would we see fewer manuscript submissions to ACS Energy Letters? I do not have a crystal ball. That could well happen, but we could end up publishing similar (or maybe even higher) numbers of high-quality papers with a higher manuscript acceptance rate (note that most journals do not disclose the manuscript acceptance rate). We might have a lower journal impact factor (JIF), (12) just as many peer journals would also likely have, but if that means that authors write fewer papers and editors and reviewers have to evaluate fewer manuscripts, and thus all of us have a lighter workload, wouldn't this be a bargain that we all should be happy to take?I thank you for reading this editorial and thinking about these issues and I welcome any debate and discussion. We at ACS Energy Letters look forward to receiving your next exciting renewable energy research work!Author InformationClick to copy section linkSection link copied!Corresponding AuthorSong Jin, Senior Editor, ACS Energy Letters, Department of Chemistry, University of Wisconsin−Madison, 1101 University Avenue, Madison, Wisconsin 53706, United States, https://orcid.org/0000-0001-8693-7010, Email: [email protected]NotesViews expressed in this editorial are those of the author and not necessarily the views of the ACS.AcknowledgmentsClick to copy section linkSection link copied!The author sincerely thanks Dr. Prashant Kamat for providing valuable feedback.ReferencesClick to copy section linkSection link copied! This article references 12 other publications. 1Hanson, M. A.; Barreiro, P. G.; Crosetto, P.; Brockington, D. The strain on scientific publishing. Arxiv 2024, arXiv.2309.15884, DOI: 10.48550/arXiv.2309.15884 Google ScholarThere is no corresponding record for this reference.2Landhuis, E. Scientific literature: Information overload. Nature 2016, 535, 457– 458, DOI: 10.1038/nj7612-457a Google Scholar2Scientific literature: Information overloadLandhuis EstherNature (2016), 535 (7612), 457-8 ISSN:. There is no expanded citation for this reference. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A280%3ADC%252BC2s3jslClug%253D%253D&md5=93979fcc6f46dd1c5d2b11bc032867fb3Geertz, C. Agricultural Involution; University of California Press, 1969.Google ScholarThere is no corresponding record for this reference.4Jin, S. Fewer Sandwich Papers, Please. ACS Energy Letters 2022, 7, 3727– 3728, DOI: 10.1021/acsenergylett.2c02197 Google Scholar4Fewer Sandwich Papers, PleaseJin, SongACS Energy Letters (2022), 7 (10), 3727-3728CODEN: AELCCP; ISSN:2380-8195. (American Chemical Society) There is no expanded citation for this reference. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB38XisFOmtr3N&md5=0fac1046e203aacd07844a053f52867c5Buriak, J. M.; Akinwande, D.; Artzi, N.; Brinker, C. J.; Burrows, C.; Chan, W. C. W.; Chen, C.; Chen, X.; Chhowalla, M.; Chi, L.; Chueh, W.; Crudden, C. M.; Di Carlo, D.; Glotzer, S. C.; Hersam, M. C.; Ho, D.; Hu, T. Y.; Huang, J.; Javey, A.; Kamat, P. V.; Kim, I.-D.; Kotov, N. A.; Lee, T. R.; Lee, Y. H.; Li, Y.; Liz-Marzán, L. M.; Mulvaney, P.; Narang, P.; Nordlander, P.; Oklu, R.; Parak, W. J.; Rogach, A. L.; Salanne, M.; Samorì, P.; Schaak, R. E.; Schanze, K. S.; Sekitani, T.; Skrabalak, S.; Sood, A. K.; Voets, I. K.; Wang, S.; Wang, S.; Wee, A. T. S.; Ye, J. Best Practices for Using AI When Writing Scientific Manuscripts. ACS Nano 2023, 17, 4091– 4093, DOI: 10.1021/acsnano.3c01544 Google Scholar5Best Practices for Using AI When Writing Scientific ManuscriptsBuriak, Jillian M.; Akinwande, Deji; Artzi, Natalie; Brinker, C. Jeffrey; Burrows, Cynthia; Chan, Warren C. W.; Chen, Chunying; Chen, Xiaodong; Chhowalla, Manish; Chi, Lifeng; Chueh, William; Crudden, Cathleen M.; Di Carlo, Dino; Glotzer, Sharon C.; Hersam, Mark C.; Ho, Dean; Hu, Tony Y.; Huang, Jiaxing; Javey, Ali; Kamat, Prashant V.; Kim, Il-Doo; Kotov, Nicholas A.; Lee, T. Randall; Lee, Young Hee; Li, Yan; Liz-Marzan, Luis M.; Mulvaney, Paul; Narang, Prineha; Nordlander, Peter; Oklu, Rahmi; Parak, Wolfgang J.; Rogach, Andrey L.; Salanne, Mathieu; Samori, Paolo; Schaak, Raymond E.; Schanze, Kirk S.; Sekitani, Tsuyoshi; Skrabalak, Sara; Sood, Ajay K.; Voets, Ilja K.; Wang, Shu; Wang, Shutao; Wee, Andrew T. S.; Ye, JinhuaACS Nano (2023), 17 (5), 4091-4093CODEN: ANCAC3; ISSN:1936-0851. (American Chemical Society) There is no expanded citation for this reference. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3sXjvVajtr8%253D&md5=471ce26b339b966476ce255670de1c586Grimaldi, G.; Ehrler, B. AI et al.: Machines Are About to Change Scientific Publishing Forever. ACS Energy Letters 2023, 8, 878– 880, DOI: 10.1021/acsenergylett.2c02828 Google ScholarThere is no corresponding record for this reference.7Buriak, J. M.; Hersam, M. C.; Kamat, P. V. Can ChatGPT and Other AI Bots Serve as Peer Reviewers?. ACS Energy Letters 2024, 9, 191– 192, DOI: 10.1021/acsenergylett.3c02586 Google ScholarThere is no corresponding record for this reference.8Chawla, D. S. Is ChatGPT corrupting peer review? Telltale words hint at AI use. Nature 2024, 628, 483– 484, DOI: 10.1038/d41586-024-01051-2 Google ScholarThere is no corresponding record for this reference.9Zheng, Z.; Zhang, O.; Borgs, C.; Chayes, Y. T.; Yaghi, O. M. ChatGPT Chemistry Assistant for Text Mining and the Prediction of MOF Synthesis. J. Am. Chem. Soc. 2023, 145 (32), 18048– 18062, DOI: 10.1021/jacs.3c05819 Google Scholar12ChatGPT Chemistry Assistant for Text Mining and the Prediction of MOF SynthesisZheng, Zhiling; Zhang, Oufan; Borgs, Christian; Chayes, Jennifer T.; Yaghi, Omar M.Journal of the American Chemical Society (2023), 145 (32), 18048-18062CODEN: JACSAT; ISSN:0002-7863. (American Chemical Society) We use prompt engineering to guide ChatGPT in the automation of text mining of metal-org. framework (MOF) synthesis conditions from diverse formats and styles of the scientific literature. This effectively mitigates ChatGPT's tendency to hallucinate information, an issue that previously made the use of large language models (LLMs) in scientific fields challenging. Our approach involves the development of a workflow implementing three different processes for text mining, programed by ChatGPT itself. All of them enable parsing, searching, filtering, classification, summarization, and data unification with different trade-offs among labor, speed, and accuracy. We deploy this system to ext. 26 257 distinct synthesis parameters pertaining to approx. 800 MOFs sourced from peer-reviewed research articles. This process incorporates our ChemPrompt Engineering strategy to instruct ChatGPT in text mining, resulting in impressive precision, recall, and F1 scores of 90-99%. Furthermore, with the data set built by text mining, we constructed a machine-learning model with over 87% accuracy in predicting MOF exptl. crystn. outcomes and preliminarily identifying important factors in MOF crystn. We also developed a reliable data-grounded MOF chatbot to answer questions about chem. reactions and synthesis procedures. Given that the process of using ChatGPT reliably mines and tabulates diverse MOF synthesis information in a unified format while using only narrative language requiring no coding expertise, we anticipate that our ChatGPT Chem. Assistant will be very useful across various other chem. subdisciplines. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3sXhs1WhtbjO&md5=3bc0454645cc6808288b5f5307610f2410https://retractionwatch.com/2023/04/05/wiley-and-hindawi-to-retract-1200-more-papers-for-compromised-peer-review/ (accessed July 23, 2024).Google ScholarThere is no corresponding record for this reference.11Hirsch, J. E. An index to quantify an individual's scientific research output. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 16569– 16572, DOI: 10.1073/pnas.0507655102 Google Scholar10An index to quantify an individual's scientific research outputHirsch, J. E.Proceedings of the National Academy of Sciences of the United States of America (2005), 102 (46), 16569-16572CODEN: PNASA6; ISSN:0027-8424. (National Academy of Sciences) I propose the index h, defined as the no. of papers with citation no. ≥h, as a useful index to characterize the scientific output of a researcher. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2MXht1Kgs7fL&md5=a90deefbe6dbaeda1de4272c308d490a12Kamat, P. V. Roller-Coaster Ride with Journal Impact Factor. ACS Energy Letters 2024, 9, 3605– 3607, DOI: 10.1021/acsenergylett.4c01709 Google ScholarThere is no corresponding record for this reference.Cited By Click to copy section linkSection link copied!This article has not yet been cited by other publications.Download PDFFiguresReferencesOpen PDF Get e-AlertsGet e-AlertsACS Energy LettersCite this: ACS Energy Lett. 2024, 9, 8, 4196–4198Click to copy citationCitation copied!https://doi.org/10.1021/acsenergylett.4c01991Published August 9, 2024 Publication History Received 23 July 2024Accepted 24 July 2024Published online 9 August 2024Published in issue 9 August 2024Copyright © 2024 American Chemical Society. This publication is available under these Terms of Use. Request reuse permissionsArticle Views11k11,719 total viewsAltmetric-Citations-Learn about these metrics closeArticle Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated.Recommended Articles FiguresReferencesFigure 1Figure 1. Researchers are overwhelmed by the number of research papers being published these days. (Source: iStock.com/aldegonde)High Resolution ImageDownload MS PowerPoint SlideReferences This article references 12 other publications. 1Hanson, M. A.; Barreiro, P. G.; Crosetto, P.; Brockington, D. The strain on scientific publishing. Arxiv 2024, arXiv.2309.15884, DOI: 10.48550/arXiv.2309.15884 There is no corresponding record for this reference.2Landhuis, E. Scientific literature: Information overload. Nature 2016, 535, 457– 458, DOI: 10.1038/nj7612-457a 2Scientific literature: Information overloadLandhuis EstherNature (2016), 535 (7612), 457-8 ISSN:. There is no expanded citation for this reference. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A280%3ADC%252BC2s3jslClug%253D%253D&md5=93979fcc6f46dd1c5d2b11bc032867fb3Geertz, C. Agricultural Involution; University of California Press, 1969.There is no corresponding record for this reference.4Jin, S. Fewer Sandwich Papers, Please. ACS Energy Letters 2022, 7, 3727– 3728, DOI: 10.1021/acsenergylett.2c02197 4Fewer Sandwich Papers, PleaseJin, SongACS Energy Letters (2022), 7 (10), 3727-3728CODEN: AELCCP; ISSN:2380-8195. (American Chemical Society) There is no expanded citation for this reference. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB38XisFOmtr3N&md5=0fac1046e203aacd07844a053f52867c5Buriak, J. M.; Akinwande, D.; Artzi, N.; Brinker, C. J.; Burrows, C.; Chan, W. C. W.; Chen, C.; Chen, X.; Chhowalla, M.; Chi, L.; Chueh, W.; Crudden, C. M.; Di Carlo, D.; Glotzer, S. C.; Hersam, M. C.; Ho, D.; Hu, T. Y.; Huang, J.; Javey, A.; Kamat, P. V.; Kim, I.-D.; Kotov, N. A.; Lee, T. R.; Lee, Y. H.; Li, Y.; Liz-Marzán, L. M.; Mulvaney, P.; Narang, P.; Nordlander, P.; Oklu, R.; Parak, W. J.; Rogach, A. L.; Salanne, M.; Samorì, P.; Schaak, R. E.; Schanze, K. S.; Sekitani, T.; Skrabalak, S.; Sood, A. K.; Voets, I. K.; Wang, S.; Wang, S.; Wee, A. T. S.; Ye, J. Best Practices for Using AI When Writing Scientific Manuscripts. ACS Nano 2023, 17, 4091– 4093, DOI: 10.1021/acsnano.3c01544 5Best Practices for Using AI When Writing Scientific ManuscriptsBuriak, Jillian M.; Akinwande, Deji; Artzi, Natalie; Brinker, C. Jeffrey; Burrows, Cynthia; Chan, Warren C. W.; Chen, Chunying; Chen, Xiaodong; Chhowalla, Manish; Chi, Lifeng; Chueh, William; Crudden, Cathleen M.; Di Carlo, Dino; Glotzer, Sharon C.; Hersam, Mark C.; Ho, Dean; Hu, Tony Y.; Huang, Jiaxing; Javey, Ali; Kamat, Prashant V.; Kim, Il-Doo; Kotov, Nicholas A.; Lee, T. Randall; Lee, Young Hee; Li, Yan; Liz-Marzan, Luis M.; Mulvaney, Paul; Narang, Prineha; Nordlander, Peter; Oklu, Rahmi; Parak, Wolfgang J.; Rogach, Andrey L.; Salanne, Mathieu; Samori, Paolo; Schaak, Raymond E.; Schanze, Kirk S.; Sekitani, Tsuyoshi; Skrabalak, Sara; Sood, Ajay K.; Voets, Ilja K.; Wang, Shu; Wang, Shutao; Wee, Andrew T. S.; Ye, JinhuaACS Nano (2023), 17 (5), 4091-4093CODEN: ANCAC3; ISSN:1936-0851. (American Chemical Society) There is no expanded citation for this reference. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3sXjvVajtr8%253D&md5=471ce26b339b966476ce255670de1c586Grimaldi, G.; Ehrler, B. AI et al.: Machines Are About to Change Scientific Publishing Forever. ACS Energy Letters 2023, 8, 878– 880, DOI: 10.1021/acsenergylett.2c02828 There is no corresponding record for this reference.7Buriak, J. M.; Hersam, M. C.; Kamat, P. V. Can ChatGPT and Other AI Bots Serve as Peer Reviewers?. ACS Energy Letters 2024, 9, 191– 192, DOI: 10.1021/acsenergylett.3c02586 There is no corresponding record for this reference.8Chawla, D. S. Is ChatGPT corrupting peer review? Telltale words hint at AI use. Nature 2024, 628, 483– 484, DOI: 10.1038/d41586-024-01051-2 There is no corresponding record for this reference.9Zheng, Z.; Zhang, O.; Borgs, C.; Chayes, Y. T.; Yaghi, O. M. ChatGPT Chemistry Assistant for Text Mining and the Prediction of MOF Synthesis. J. Am. Chem. Soc. 2023, 145 (32), 18048– 18062, DOI: 10.1021/jacs.3c05819 12ChatGPT Chemistry Assistant for Text Mining and the Prediction of MOF SynthesisZheng, Zhiling; Zhang, Oufan; Borgs, Christian; Chayes, Jennifer T.; Yaghi, Omar M.Journal of the American Chemical Society (2023), 145 (32), 18048-18062CODEN: JACSAT; ISSN:0002-7863. (American Chemical Society) We use prompt engineering to guide ChatGPT in the automation of text mining of metal-org. framework (MOF) synthesis conditions from diverse formats and styles of the scientific literature. This effectively mitigates ChatGPT's tendency to hallucinate information, an issue that previously made the use of large language models (LLMs) in scientific fields challenging. Our approach involves the development of a workflow implementing three different processes for text mining, programed by ChatGPT itself. All of them enable parsing, searching, filtering, classification, summarization, and data unification with different trade-offs among labor, speed, and accuracy. We deploy this system to ext. 26 257 distinct synthesis parameters pertaining to approx. 800 MOFs sourced from peer-reviewed research articles. This process incorporates our ChemPrompt Engineering strategy to instruct ChatGPT in text mining, resulting in impressive precision, recall, and F1 scores of 90-99%. Furthermore, with the data set built by text mining, we constructed a machine-learning model with over 87% accuracy in predicting MOF exptl. crystn. outcomes and preliminarily identifying important factors in MOF crystn. We also developed a reliable data-grounded MOF chatbot to answer questions about chem. reactions and synthesis procedures. Given that the process of using ChatGPT reliably mines and tabulates diverse MOF synthesis information in a unified format while using only narrative language requiring no coding expertise, we anticipate that our ChatGPT Chem. Assistant will be very useful across various other chem. subdisciplines. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3sXhs1WhtbjO&md5=3bc0454645cc6808288b5f5307610f2410https://retractionwatch.com/2023/04/05/wiley-and-hindawi-to-retract-1200-more-papers-for-compromised-peer-review/ (accessed July 23, 2024).There is no corresponding record for this reference.11Hirsch, J. E. An index to quantify an individual's scientific research output. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 16569– 16572, DOI: 10.1073/pnas.0507655102 10An index to quantify an individual's scientific research outputHirsch, J. E.Proceedings of the National Academy of Sciences of the United States of America (2005), 102 (46), 16569-16572CODEN: PNASA6; ISSN:0027-8424. (National Academy of Sciences) I propose the index h, defined as the no. of papers with citation no. ≥h, as a useful index to characterize the scientific output of a researcher. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2MXht1Kgs7fL&md5=a90deefbe6dbaeda1de4272c308d490a12Kamat, P. V. Roller-Coaster Ride with Journal Impact Factor. ACS Energy Letters 2024, 9, 3605– 3607, DOI: 10.1021/acsenergylett.4c01709 There is no corresponding record for this reference.