Explore the vast range of titles available.

Computational social science is more than just large repositories of digital data and the computational methods needed to construct and analyse them. It also represents a convergence of different fields with different ways of thinking about and doing science. The goal of this Perspective is to provide some clarity around how these approaches differ from one another and to propose how they might be productively integrated. Towards this end we make two contributions. The first is a schema for thinking about research activities along two dimensions—the extent to which work is explanatory, focusing on identifying and estimating causal effects, and the degree of consideration given to testing predictions of outcomes—and how these two priorities can complement, rather than compete with, one another. Our second contribution is to advocate that computational social scientists devote more attention to combining prediction and explanation, which we call integrative modelling, and to outline some practical suggestions for realizing this goal. The combination of computational and social sciences requires the integration of explanatory and predictive approaches into ‘integrative modelling’, according to Hofman and colleagues.

Zipf-like distributions characterize a wide set of phenomena in physics, biology, economics, and social sciences. In human activities, Zipf's law describes, for example, the frequency of appearance of words in a text or the purchase types in shopping patterns. In the latter, the uneven distribution of transaction types is bound with the temporal sequences of purchases of individual choices. In this work, we define a framework using a text compression technique on the sequences of credit card purchases to detect ubiquitous patterns of collective behavior. Clustering the consumers by their similarity in purchase sequences, we detect five consumer groups. Remarkably, post checking, individuals in each group are also similar in their age, total expenditure, gender, and the diversity of their social and mobility networks extracted from their mobile phone records. By properly deconstructing transaction data with Zipf-like distributions, this method uncovers sets of significant sequences that reveal insights on collective human behavior. Digital traces of our lives have the potential to allow insights into collective behaviors. Here, the authors cluster consumers by their credit card purchase sequences and discover five distinct groups, within which individuals also share similar mobility and demographic attributes.

We examined the long-term impact of coauthorship with established, highly-cited scientists on the careers of junior researchers in four scientific disciplines. Here, using matched pair analysis, we find that junior researchers who coauthor work with top scientists enjoy a persistent competitive advantage throughout the rest of their careers, compared to peers with similar early career profiles but without top coauthors. Such early coauthorship predicts a higher probability of repeatedly coauthoring work with top-cited scientists, and, ultimately, a higher probability of becoming one. Junior researchers affiliated with less prestigious institutions show the most benefits from coauthorship with a top scientist. As a consequence, we argue that such institutions may hold vast amounts of untapped potential, which may be realised by improving access to top scientists. By examining publication records of scientists from four disciplines, the authors show that coauthoring a paper with a top-cited scientist early in one's career predicts lasting increases in career success, especially for researchers affiliated with less prestigious institutions.

ChatGPT is a societally impactful artificial intelligence tool with millions of users and integration into products such as Bing. However, the emergence of jailbreak attacks notably threatens its responsible and secure use. Jailbreak attacks use adversarial prompts to bypass ChatGPT’s ethics safeguards and engender harmful responses. This paper investigates the severe yet under-explored problems created by jailbreaks as well as potential defensive techniques. We introduce a jailbreak dataset with various types of jailbreak prompts and malicious instructions. We draw inspiration from the psychological concept of self-reminders and further propose a simple yet effective defence technique called system-mode self-reminder. This technique encapsulates the user’s query in a system prompt that reminds ChatGPT to respond responsibly. Experimental results demonstrate that self-reminders significantly reduce the success rate of jailbreak attacks against ChatGPT from 67.21% to 19.34%. Our work systematically documents the threats posed by jailbreak attacks, introduces and analyses a dataset for evaluating defensive interventions and proposes the psychologically inspired self-reminder technique that can efficiently and effectively mitigate against jailbreaks without further training. Interest in using large language models such as ChatGPT has grown rapidly, but concerns about safe and responsible use have emerged, in part because adversarial prompts can bypass existing safeguards with so-called jailbreak attacks. Wu et al. build a dataset of various types of jailbreak attack prompt and demonstrate a simple but effective technique to counter these attacks by encapsulating users’ prompts in another standard prompt that reminds ChatGPT to respond responsibly.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) continued to mutate and spread in 2022 despite the introduction of safe, effective vaccines and medications. Vaccine hesitancy remains substantial, fueled in part by misinformation. Our third study of Coronavirus Disease 2019 (COVID-19) vaccine hesitancy among 23,000 respondents in 23 countries (Brazil, Canada, China, Ecuador, France, Germany, Ghana, India, Italy, Kenya, Mexico, Nigeria, Peru, Poland, Russia, Singapore, South Africa, South Korea, Spain, Sweden, Turkey, the United Kingdom and the United States), surveyed from 29 June to 10 July 2022, found willingness to accept vaccination at 79.1%, up 5.2% from June 2021. Hesitancy increased in eight countries, however, ranging from 1.0% (United Kingdom) to 21.1% (South Africa). Almost one in eight (12.1%) vaccinated respondents are hesitant about booster doses. Overall support for vaccinating children under 18 years of age increased slightly but declined among parents who were personally hesitant. Almost two in five (38.6%) respondents reported paying less attention to new COVID-19 information than previously, and support for vaccination mandates decreased. Almost a quarter (24%) of those who became ill reported taking medications to combat COVID-19 symptoms. Vaccination remains a cornerstone of the COVID-19 pandemic response, but broad public support remains elusive. These data can be used by health system decisionmakers, practitioners, advocates and researchers to address COVID-19 vaccine hesitancy more effectively. In the third annual iteration of a survey assessing vaccine confidence in 23 high-income, middle-income and low-income countries, although results were heteregeous across countries, overall willingness to accept a COVID-19 vaccine was found to have increased.

Air pollution kills nearly 1 million people per year in China. In response, the Chinese government implemented the Air Pollution Prevention and Control Action Plan (APPCAP) from 2013 to 2017 which had a significant impact on reducing PM 2.5 concentration. However, the health benefits of the APPCAP are not well understood. Here we examine the spatiotemporal dynamics of annual deaths attributable to PM 2.5 pollution (DAPP) in China and the contribution from the APPCAP using decomposition analysis. Despite a 36.1% increase in DAPP from 2000 to 2017, The APPCAP-induced improvement in air quality achieved substantial health benefits, with the DAPP in 2017 reduced by 64 thousand (6.8%) compared to 2013. However, the policy is unlikely to result in further major reductions in DAPP and more ambitious policies are required to reduce the health impacts of air pollution by 2030 and meet the United Nation’s Sustainable Development Goal 3. Chinese government has implemented the air pollution control measure-the Air Pollution Prevention and Control Action Plan in 2013, whose effects have not been fully studied. Here the authors show that from 2013 to 2017, the plan has achieved substantial public health benefits.

Scientists are enthusiastically imagining ways in which artificial intelligence (AI) tools might improve research. Why are AI tools so attractive and what are the risks of implementing them across the research pipeline? Here we develop a taxonomy of scientists’ visions for AI, observing that their appeal comes from promises to improve productivity and objectivity by overcoming human shortcomings. But proposed AI solutions can also exploit our cognitive limitations, making us vulnerable to illusions of understanding in which we believe we understand more about the world than we actually do. Such illusions obscure the scientific community’s ability to see the formation of scientific monocultures, in which some types of methods, questions and viewpoints come to dominate alternative approaches, making science less innovative and more vulnerable to errors. The proliferation of AI tools in science risks introducing a phase of scientific enquiry in which we produce more but understand less. By analysing the appeal of these tools, we provide a framework for advancing discussions of responsible knowledge production in the age of AI. The proliferation of artificial intelligence tools in scientific research risks creating illusions of understanding, where scientists believe they understand more about the world than they actually do.

Widespread acceptance of COVID-19 vaccines is crucial for achieving sufficient immunization coverage to end the global pandemic, yet few studies have investigated COVID-19 vaccination attitudes in lower-income countries, where large-scale vaccination is just beginning. We analyze COVID-19 vaccine acceptance across 15 survey samples covering 10 low- and middle-income countries (LMICs) in Asia, Africa and South America, Russia (an upper-middle-income country) and the United States, including a total of 44,260 individuals. We find considerably higher willingness to take a COVID-19 vaccine in our LMIC samples (mean 80.3%; median 78%; range 30.1 percentage points) compared with the United States (mean 64.6%) and Russia (mean 30.4%). Vaccine acceptance in LMICs is primarily explained by an interest in personal protection against COVID-19, while concern about side effects is the most common reason for hesitancy. Health workers are the most trusted sources of guidance about COVID-19 vaccines. Evidence from this sample of LMICs suggests that prioritizing vaccine distribution to the Global South should yield high returns in advancing global immunization coverage. Vaccination campaigns should focus on translating the high levels of stated acceptance into actual uptake. Messages highlighting vaccine efficacy and safety, delivered by healthcare workers, could be effective for addressing any remaining hesitancy in the analyzed LMICs.

Humanity has become a dominant force in shaping the face of Earth 1 – 9 . An emerging question is how the overall material output of human activities compares to the overall natural biomass. Here we quantify the human-made mass, referred to as ‘anthropogenic mass’, and compare it to the overall living biomass on Earth, which currently equals approximately 1.1 teratonnes 10 , 11 . We find that Earth is exactly at the crossover point; in the year 2020 (± 6), the anthropogenic mass, which has recently doubled roughly every 20 years, will surpass all global living biomass. On average, for each person on the globe, anthropogenic mass equal to more than his or her bodyweight is produced every week. This quantification of the human enterprise gives a mass-based quantitative and symbolic characterization of the human-induced epoch of the Anthropocene. Estimates of global total biomass (the mass of all living things) and anthopogenic mass (the mass embedded in inanimate objects made by humans) over time show that we are roughly at the timepoint when anthropogenic mass exceeds total biomass.

Climate change and political polarization are two of the twenty-first century’s critical socio-political issues. Here we investigate their intersection by studying the discussion around the United Nations Conference of the Parties on Climate Change (COP) using Twitter data from 2014 to 2021. First, we reveal a large increase in ideological polarization during COP26, following low polarization between COP20 and COP25. Second, we show that this increase is driven by growing right-wing activity, a fourfold increase since COP21 relative to pro-climate groups. Finally, we identify a broad range of ‘climate contrarian’ views during COP26, emphasizing the theme of political hypocrisy as a topic of cross-ideological appeal; contrarian views and accusations of hypocrisy have become key themes in the Twitter climate discussion since 2019. With future climate action reliant on negotiations at COP27 and beyond, our results highlight the importance of monitoring polarization and its impacts in the public climate discourse.Polarization and the resulting political deadlock have become key barriers to more ambitious climate action. Using Twitter data between Conferences of the Parties, this research identifies a trend of increasing polarization driven by growing right-wing activity alongside accusations of political hypocrisy.