Researchers who started their careers in the previous millennium may recall the days before the Internet liberated them from the task of sorting through the exceedingly thin pages of Chemical Abstracts, hauling obscure journals out of the library stacks and stapling together copies of papers. More recently, those same computers have started to make recommendations about papers we might want to read. They also edit our grammar and spelling and suggest polite responses to routine e-mails. In the coming decades, it seems likely that artificial intelligence will continue to alter the way that researchers spend their time. I expect that the last thing the machines will learn to do is select the best papers of the year. A computer may be able to determine when a manuscript exceeds some minimal threshold with respect to presentation, data quality, and analysis, but it will be difficult for an algorithm to identify aspects of the work that separate a truly exceptional paper from one that is good enough to publish.
Decisions about excellence in research still require the judgment of experts who have years of research experience, diverse skill sets, and a global perspective. For this reason, when it is time to pick the best papers of the year, Environmental Science& Technology
relies upon our Editorial Advisory Board
. Each year we follow a similar process. The Board reads and ranks approximately 100 papers that our Associate Editors had identified as exceptional contributions. From among the top-ranked papers, I choose best papers in the categories of Environmental Science, Environmental Technology, Environmental Policy, and Features. Ranking all of those papers was a tough job, and we are particularly grateful for the tireless efforts of Professor Patricia Holden (UC Santa Barbara), the chair of the selection committee.
If I were asked to advise a computer scientist about how to create a program that could choose a best paper, I would tell them to look for those papers where researchers report on a series of unexpected findings. For example, our top science paper
puts together the cause-and-effect relationships between the neonicotinoid pesticides that have become a staple of modern agriculture and the increase in resistant insects in a heavily farmed region. I would tell the computer scientist to pay attention to research that builds upon recent developments in rapidly evolving fields and applies them to conditions that are encountered in the environment. Our top technology paper
is a perfect example of how developments in materials science and electrochemistry are allowing researchers to create electrodes that can turn the dream of chemical reduction of nitrate into a reality. Finally, I would advise her to seek those papers that had implications for public policy and future decisions about where additional research should be directed. This year, I would suggest that our computer scientist use all three of our top Features to train their computers how to recognize insightful, policy-relevant research.
We are grateful to the exceptional researchers whose papers we recognize with these awards. We also appreciate the thousands of other authors whose research is reported in the journal. And finally, we are appreciative of the effort of our Editorial Advisory Board. The machines are not going to replace you anytime soon!