The fear of unintended consequences is frequently used to argue against conservation interventions that range from climate engineering, to genetic editing of imperiled species, to actions as seemingly mundane as using seeds from non‐local sources in restoration projects. There is no denying that unintended consequences are real and worthy of concern. Indeed, environmental textbooks are filled with descriptions of past interventions gone awry (such as the introduction of cane toads to Australia for biocontrol, the impacts of long‐term wildfire suppression, or the use of DDT to control insect pests). However, there are also numerous counterexamples of interventions turning out as planned (for instance, barging salmon smolts around Snake River dams and human‐assisted hybridization rescuing the depleted gene pool of Florida panthers). For actions under consideration, the question is how to weigh the possible unintended consequences versus the highly likely intended benefits.

A June 2020 workshop (https://reviv​erest​ore.org/inten​ded-conse​quences), organized by Revive & Restore, assembled an international group of conservationists (including wildlife biologists, restoration scientists, geneticists, ethicists, and social scientists) to re‐examine the precautionary principle and its associated focus on unintended and unanticipated consequences. Two observations make apparent the need for this reassessment. First, accelerating anthropogenic climate change and the expanding human footprint create ever‐greater urgency for actions that could avert disasters or prevent human‐driven extinctions. With most rivers dammed, a nitrogen cycle dominated by human alterations, and such severe global warming that within 50 years as many as one in three humans could be forced to migrate in search of a habitable environment (https://nyti.ms/2E5a0Wi), conservationists do not have the luxury of “doing nothing” out of fear of unintended consequences. Caution is prudent, but paralysis is unconscionable. Second, the science of risk assessment has advanced so that, although it is impossible to eliminate uncertainty, the likelihood of horrific ecological surprises is much less now than in previous decades. Ecologists today better understand ecosystems and indirect effects than they did in the 1960s and 1970s when several well‐intended, but ill‐fated, introductions were conducted (eg introducing the American red squirrel to Newfoundland to augment the diet of pine martens; introducing the seed weevil Rhinocyllus conicus to control exotic thistles in North America). Lessons learned with each intervention reduce the chance for future errors. Not only is risk assessment improving, but some of the tools for intervention are becoming refined. For example, the genomic and phenotypic changes that result from gene editing are much more precise than those wrought by more widely accepted techniques like hybridization and mutagenesis.

Obviously, the answer is not to blithely ignore unintended consequences and adopt an “anything goes” attitude. However, it is time to recalibrate our traditional cautionary approach to environmental decision making with fine‐tuning in four dimensions. First, the intended consequences of proposed interventions must carry more weight in analyses. Consider the public debate over transgenic Bt crops. The intended benefit of reducing the application of broadly toxic insecticides has largely been overshadowed by fears of environmental harms, for which there is no convincing evidence. Second, scientists should avoid being overly influenced by examples of disaster from generations ago and give more credit to recently accumulated empirical evidence. From reintroductions to genetic rescues, the vast majority of contemporary environmental interventions have produced their intended positive outcomes, yet a few historical cases of problematic outcomes continue to dominate public perception. Moreover, all actions, including inaction, entail the potential for unintended consequences. For instance, growing evidence shows that protected area creation – an intervention largely embraced by conservationists – typically displaces, rather than curtails, environmental harms and can prove counterproductive if local communities are alienated. Third, risk assessment relies on tools ranging from controlled experiments and practical experience to models and simulations. All of these tools help characterize risk, but empirical data should be far more reassuring than theoretical, but untested, models. Thirty years of globally widespread Bt crops with no ill‐effects ought to inspire confidence, whereas speculative interventions such as solar geoengineering warrant greater precaution. Lastly, and most importantly, scientists need to recognize that the worst unintended consequences may not be environmental or ecological, but rather social. A carbon tax might be a great way to reduce greenhouse‐gas emissions and a protected area may secure a remnant population of a declining species, but such actions disproportionately impose costs on marginalized human communities. A wider variety of voices must be invited to weigh in on what intended outcomes are desired, and how best to achieve them. While risk reduction efforts now better recognize and minimize unintended environmental harms, much work remains to address unintended social and cultural consequences.

人们常常担心产生意想不到的后果，以反对采取保护措施，这些干预措施包括气候工程，对受威胁物种的基因编辑，以及在恢复项目中使用来自非本地来源的种子的看似平凡的行动。无可否认，意想不到的后果是真实的，值得关注。确实，环境教科书中充斥着对过去的干预措施失误的描述（例如，将甘蔗蟾蜍引入澳大利亚进行生物防治，长期抑制野火的影响或使用滴滴涕控制虫害）。但是，也有许多按计划实施的干预措施的反例（例如，在蛇河水坝附近拦住鲑鱼鲑鱼，人工辅助杂交以拯救佛罗里达黑豹的枯竭基因库）。可能的意外结果与极有可能的预期收益。

由Revive＆Restore组织的2020年6月的研讨会（https://revivererestore.org/intended-consequences）召集了一个国际保护主义者小组（包括野生生物生物学家，恢复科学家，遗传学家，伦理学家，和社会科学家）重新审视预防原则及其对意外后果的关注。有两个观察结果表明需要进行重新评估。首先，人为气候变化的加速发展和人类足迹的扩大给采取紧急行动的可能性越来越大，这些行动可以避免灾难或防止人类灭绝。在大多数河流被筑坝的情况下，氮循环主要由人类改变以及如此严重的全球变暖导致，在50年内，多达三分之一的人类可能被迫迁移以寻找适宜的环境（https：//nyti.ms/2E5a0Wi），出于对意外后果的恐惧，环保主义者没有“无所作为”的奢望。谨慎谨慎，但瘫痪是不合情理的。其次，风险评估的科学已经发展，因此尽管不可能消除不确定性，但如今发生可怕的生态意外的可能性比过去几十年要少得多。今天的生态学家比1960年代和1970年代对生态系统及其间接影响的了解要好得多，当时曾进行了几次有意但不幸的介绍（例如，将美国红松鼠引入纽芬兰以增加松貂的饮食；引入种子）。象鼻虫 风险评估的科学不断发展，因此尽管不可能消除不确定性，但如今发生可怕的生态意外的可能性比过去几十年要少得多。今天的生态学家比1960年代和1970年代对生态系统及其间接影响的理解要好得多，当时曾进行了几次有意但不幸的介绍（例如，将美国红松鼠引入纽芬兰以增加松貂的饮食；引入种子）。象鼻虫 风险评估的科学不断发展，因此尽管不可能消除不确定性，但如今发生可怕的生态意外的可能性比过去几十年要少得多。今天的生态学家比1960年代和1970年代对生态系统及其间接影响的了解要好得多，当时曾进行了几次有意但不幸的介绍（例如，将美国红松鼠引入纽芬兰以增加松貂的饮食；引入种子）。象鼻虫Rhinocyllus conicus控制北美的外来蓟。每次干预都会吸取教训，从而减少了将来出错的机会。不仅风险评估有所改善，而且一些干预工具也在不断完善。例如，基因编辑产生的基因组和表型变化比杂交和诱变等更广泛接受的技术所产生的变化更为精确。

显然，答案是不要无理地忽略意想不到的后果，并采取“随心所欲”的态度。但是，现在该是时候通过四个方面的微调来重新调整我们传统的环境决策预警方法了。首先，拟议干预措施的预期结果必须在分析中发挥更大的作用。考虑关于转基因Bt作物的公开辩论。减少对广泛毒性杀虫剂的使用的预期利益在很大程度上被对环境危害的担忧所掩盖，对此没有令人信服的证据。其次，科学家应避免受到世代前代灾难事例的过度影响，并应更多地归功于最近积累的经验证据。从重新引入到基因拯救，绝大多数当代环境干预措施已产生了预期的积极成果，但历史上有问题的案例仍继续在公众认知中占主导地位。此外，所有行动，包括不作为，都可能带来意想不到的后果。例如，越来越多的证据表明，保护区的建立是保护主义者广泛接受的一项干预措施，通常会取代而不是减少环境危害，如果疏远当地社区，可能会适得其反。第三，风险评估依赖于从受控实验和实践经验到模型和仿真的工具。所有这些工具都有助于表征风险，但经验数据应比理论模型（未经测试）更加放心。在全球范围内种植30年无不良影响的Bt作物应该激发人们的信心，而诸如太阳能地球工程等投机性干预措施则需要采取更大的预防措施。最后，也是最重要的是，科学家需要认识到，最严重的意外后果可能不是环境或生态的后果，而是社会的后果。碳税可能是减少温室气体排放的一种好方法，保护区可能会使残留的物种减少，但这种行动不成比例地给边缘化的人类社区带来了代价。必须邀请各种各样的声音来权衡期望的预期结果以及如何最好地实现这些结果。现在，尽管减少风险的努力可以更好地认识到并最大程度减少意外的环境危害，