Epidemic: Overall, this study lends support to the idea that smartphones can be beneficial for individuals, particularly during times when face-to-face interaction is limited

Did smartphones enhance or diminish well-being during the COVID-19 pandemic? Jennifer L. Heyman and Kostadin Kushlev. Front. Psychol., March 13 2023, Volume 14 - 2023 | https://doi.org/10.3389/fpsyg.2023.1094196

Introduction: As smartphones have become increasingly integrated into people’s lives, researchers have attempted to answer whether they are beneficial or detrimental to well-being. Of particular interest to the current study is the role that smartphones played during the first year of the COVID-19 Pandemic.

Methods: In an intensive longitudinal study, we explore how varying uses of smartphones relate to well-being using the Displacement-Interference-Complementarity framework.

Results: Consistent with pre-pandemic research, we show that people felt better, calmer, and more energetic when they used their phones more for complementary purposes (i.e., to access information, entertainment, and connection not otherwise available). In contrast to most pre-pandemic research, however, we find no evidence that any type of phone use predicted lower well-being during the pandemic.

Discussion: Overall, this study lends support to the idea that smartphones can be beneficial for individuals, particularly during times when face-to-face interaction is limited.

Discussion

We find that in a time of high social distancing during the COVID-19 pandemic, people reaped the benefits of phone use for well-being without incurring the costs associated with phone use in pre-pandemic research. Specifically, consistent with pre-pandemic research (e.g., Kushlev et al., 2017), we find that people who used their phones in a complementary way—to access information, entertainment, and connection not otherwise available—felt better, calmer, and more energetic. Furthermore, we show that the same individuals felt better, calmer and more energetic on days when they used their phones for complementary purposes. Pre-pandemic research also shows, however, that phone use often undermines well-being, especially when it displaces (Lanaj et al., 2014; Hughes and Burke, 2018) or interferes with other activities (Dwyer et al., 2018; Kushlev and Dunn, 2019). In contrast, we found no evidence that phone interference or displacement predicted lower well-being during the initial stages of the pandemic. Thus, though the pre-pandemic literature has generally linked phone use and screentime with poorer well-being (Twenge and Campbell, 2019), we find that phone use during the pandemic was associated with higher, not lower well-being.

In line with previous research, phone complementarity was related to higher levels of well-being. That is, the greater affordance to information and opportunities provided by a phone was related to people having better moods, feeling calmer, and feeling more energetic. The ease of access to information and opportunities may have become even more important during the COVID-19 Pandemic when face-to-face social contact was severely limited, which significantly increased people’s level of stress (Halliburton et al., 2021). Therefore, using one’s phone to maintain existing relationships and gain access to information may have facilitated in maintaining some semblance of pre-pandemic life, thus predicting higher well-being.

People typically feel worse when their phone use displaces activities critical for well-being, such as sleep (Lanaj et al., 2014). We find little evidence that phone displacement undermined well-being during the pandemic. This may be because there were fewer positive activities that phone use could displace during the pandemic when social activities and events were discouraged. Presumably, however, people needed just as much sleep during the pandemic as they did pre-pandemic. As lockdowns disrupted routines, sleep–wake cycles were delayed during the pandemic (Sinha et al., 2020). Thus, in the relative lack of routine during the pandemic, phone use may have been less likely to displace sleep. Finally, as the pandemic introduced new stressors, phone displacement might have been beneficial for well-being by displacing more stressful activities (Kushlev and Leitao, 2020) and introducing a welcome source of distraction (Sheppes and Meiran, 2007; Quoidbach et al., 2010).

In contrast to pre-pandemic research, we found no evidence that phone interference predicted lower well-being. Just as with displacement, this lack of effect may be due to the relative lack of rewarding activities associated with social distancing. Indeed, most previous research on the interference effects of phones has shown that phones decrease well-being precisely by interfering with face-to-face social interactions (Dwyer et al., 2018; Kushlev and Dunn, 2019). In addition, during the COVID-19 Pandemic, phones may have also interfered with activities harmful to well-being, such as rumination. Overall, then, though null findings should be interpreted with caution, our evidence suggests that phone use may not have been as harmful during the COVID-19 Pandemic.

Our findings were generally consistent with the Displacement–Interreference–Complementarity Framework: During a time of limited rewarding activities, complementary phone use continued to predict higher well-being, whereas well-documented phone interference and displacement effects were absent. According to the framework, however, at higher levels of social distancing, phone complementarity effects should have been stronger and phone displacement and interference effects should have been weaker. But we found little evidence that these effects depended on how much people socially distanced. Other research during the pandemic, however, showed that the benefits of online social interactions for well-being were greater when social distancing measures were more extreme (Marinucci et al., 2022). Specifically, online social interactions predicted lower distress only during the severe isolation stage in Italy that included prohibiting people from leaving their homes except for work and urgent health reasons. The social distancing measures that our participants in the United States experienced were much milder in comparison and participants, on average, reported high but not extreme levels of practicing social distancing (M = 3.12 on a scale from 1–not at all to 4–completely). Relatedly, people in our sample did not differ much in the extent to which they practiced social distancing, potentially preventing us from detecting moderating effects. Indeed, the extent to which people varied in their social distancing practices was low in this sample at both the within (SD = 0.58) and between (SD = 0.71) person levels.

This study had several important limitations that should be discussed. First, participants self-report on their levels of phone displacement, interference, and complementarity. However, people tend to misestimate the extent to which they use their phones. Future research should use more objective techniques, such as phone tracking, or peer reports in accordance with self-reports to gain a better understanding of how people are using their phones and the extent to which it relates to well-being. In addition, we used ad hoc measures of displacement, interference, and complementarity. Though theoretically justified, it is important for future research to develop validated measures of these constructs. For example, we measured phone displacement as the amount of time people spent on their phones in bed, the extent to which they used their phones more than they wanted to, and their total screentime. This crude measure of displacement fails to distinguish between screen time that displaces positive versus negative activities. As such, future research should utilize more precise measures of phone displacement, perhaps by explicitly asking people if they chose to use their phones over partaking in specific other activities. Furthermore, this study was conducted solely in the United States. However, other countries tend to use their phones in different ways (Langer et al., 2017) and have had different responses to the COVID-19 Pandemic (Kennedy et al., 2020). Therefore, future research should collect a more diverse sample to improve the generalizability of these results.

In sum, there is consistent evidence to suggest that using one’s phone for complementary purposes is associated with increases in well-being, as indicated by better mood, feeling calmer, and feeling more energetic, whereas spending more time on one’s phone and reporting that one’s phone interferes with daily life are generally not significantly associated with feeling good, calm, or energetic. Furthermore, we do not find consistent evidence that social distancing influences these associations. This study highlights the idea that phone use can be beneficial to individual’s well-being if it is used to complement their existing experiences.

Prevalence of Intoxicating Substance Use Before or During Sex Among Young Adults: Young adults show a high prevalence of the use intoxicating substances before or during sex, with women competing with men in every respect

Prevalence of Intoxicating Substance Use Before or During Sex Among Young Adults: A Systematic Review and Meta-Analysis. María Isabel Gómez-Núñez, Cristian Molla-Esparza, Natalia Gandia Carbonell & Laura Badenes Ribera. Archives of Sexual Behavior, Mar 10 2023. https://link.springer.com/article/10.1007/s10508-023-02572-z

Abstract: Drug use before or during sex is a high-risk sexual behavior associated with adverse health risks and outcomes, such as increasing the likelihood of overdoses and of acquiring sexually-transmitted diseases. This systematic review and meta-analysis of three scientific databases examined the prevalence of the use of intoxicating substances, those tending to excite or stupefy the user on a psychoactive level, before or during sex, among young adults (18–29 years old). A total of 55 unique empirical studies met the inclusion criteria (48,145 individuals; 39% males), were assessed for risk of bias using the tools of Hoy et al. (2012), and were analyzed via a generalized linear mixed-effects model. The results produced a global mean prevalence of this sexual risk behavior of 36.98% (95% CI: 28.28%, 46.63%). Nonetheless, significant differences were identified between different intoxicating substances, with the use of alcohol (35.10%; 95% CI: 27.68%, 43.31%), marijuana (27.80%; 95% CI: 18.24%, 39.92%), and ecstasy (20.90%; 95% CI: 14.34%, 29.45%) significantly more prevalent than that of cocaine (4.32%; 95% CI: 3.64%, 5.11%), heroin (.67%; 95% CI: .09%, 4.65%), methamphetamine (7.10%; 95% CI: 4.57%, 10.88%), and GHB (6.55%; 95% CI: 4.21%, 10.05%). Moderator analyses showed that the prevalence of alcohol use before or during sex differed according to geographical sample origin, and increased as the proportion of ethnic whites in samples increased. The remaining demographic (e.g., gender, age, reference population), sexual (e.g., sexual orientation, sexual activity), health (e.g., drug consumption, STI/STD status), methodological (e.g., sampling technique), and measurement (e.g., timeframe) variables that were examined did not moderate prevalence estimates. Implications for sexual development interventions were discussed.

Discussion

This study provides a systematic review and meta-estimate of the prevalence of the use of intoxicating substances before or during sex among young adults. To the best of our knowledge, this is the first study that specifically examines the prevalence of this behavior in this age grouping. Previous literature has mainly focused on analyzing SDU in MSM communities (Edmundson et al., 2018; Guerra et al., 2020; Íncera-Fernández et al., 2021; Lafortune et al., 2021; Maxwell et al., 2019), which makes it difficult to assess comparability.

Our results have revealed a high global prevalence of the use of intoxicating substances before or during sex, that is, 28–47% of young adults. This high prevalence estimate concurs with those reported in the abovementioned review studies that specifically focused on examining SDU among MSM (Edmundson et al., 2018; Maxwell et al., 2019). Nonetheless, comparisons should be made with caution, considering the differences in the meta-analytical samples. The results obtained in this research also indicate that the prevalence of using an intoxicating substance before or during sex varies significantly according to the specific substance used. The consumption of alcohol, marijuana, and ecstasy had significantly higher prevalences than those referring to the use of cocaine, heroin, methamphetamine, and GHB, perhaps due to differences in their acceptance, accessibility, cost, addictive potential, and short- and medium-term health and social consequences (Bourne et al., 2014; Graupensperger et al., 2021; Jackson et al., 2021; Rosińska et al., 2018). It is also likely that higher prevalence rates and the decision to use one or another substance depends on the particular effects it causes during sex. Alcohol may be used to become disinhibited before sex (Herbenick et al., 2021), even though it may make it more difficult to reach orgasm and reduce quality of sex (George, 2019; Palamar et al., 2018a). Ecstasy, on the other hand, may affect sexual experience the most, for example, by prolonging erection duration (Coyer et al., 2022). Future research should go more deeply into the reasons that lead to the choice of one or another drug, considering the sexual act as a complex process (e.g., from flirtation right through to after-sex behaviors), distinguishing the type and quantity of substances consumed, and the order in which they are consumed, if more than one is involved. The decision to consume one drug or another before or during sex may also be associated with individual characteristics (e.g., conception of sexuality and attitudes towards substance use for pleasure-seeking) and demographic characteristics (e.g., age and work-related aspects, including income level) and, for example, older cohorts may use higher-cost and harder-to-access substances (e.g., methamphetamine and cocaine). Regarding the geographical origin of samples, it was found that the estimated prevalence of alcohol use before or during sex varied significantly and was lowest on the African continent. Geographical comparisons pose a particular challenge due to significant continent-level, country-level, region-level, and population subgroup differences. Though this study’s comparisons are made for descriptive purposes only, several multi-level societal and individual factors may explain the observed differences in alcohol consumption patterns, including economic development, cultural aspects, such as cultural norms and religious beliefs, and societal and political aspects, such as rules concerning drug access and use, and the effectiveness of alcohol control policies (Addo et al., 2018). Future studies may further explore such factors in order to gain more insight into the driving forces behind such variations.

The meta-regression analyses of this study additionally indicate that the prevalence of using intoxicating substances before or during sex differs significantly as a function of the ethnic composition of samples. When considering alcohol, the prevalence was significantly higher in studies with a higher proportion of ethnic whites in their samples. This is in agreement with a previous meta-analysis suggesting that risky sex and drug-use behaviors were reported most frequently by studies that sampled more ethnic whites (Cunningham et al., 2017). However, to our knowledge, reasons for such differences have not been examined yet in the literature. Previous empirical studies have suggested that individual developmental trajectories and socio-cultural and economic factors may help explain ethnic trends and differences in drug use among adolescents and young adults (Chen & Jacobson, 2012; Evans et al., 2017; Vaughn et al., 2018), with white youths having higher rates of illicit drug and alcohol abuse (Johnston et al., 2019; Jones et al., 2020). Future research might explore such aspects in order to improve our understanding of protective and risk exposure factors that may elucidate potential ethnic differences in the use of intoxicating substances before or during sex. However, when considering the global prevalence, a significant inverse relationship is seen. This opposing result is not directly interpretable, since there may be compensatory effects deriving from raw measures that do not distinguish between substances (Cunningham et al., 2017). Other demographic and sexual variables, such as gender, age or sexual orientation, did not show statistically significant effects on the prevalence of intoxicating substance use before or during sex. Regarding gender, for example, the findings of our meta-analysis concur with recent empirical studies suggesting that women present rates of alcohol or other substance use similar to that of men (Ford et al., 2021b; McKetta et al., 2022), although these studies did not focus on consumption before or during sexual activity. Developmental and sociocultural factors may explain the non-difference concerning these variables. For example, young adulthood is a developmental period characterized by the reinforcement of one’s own sexual identity and orientation, in which drugs and sex may be taken as a means of experiencing a variety of sexual practices in common pleasure-seeking (Arnett, 2005, 2007, 2014), regardless of the abovementioned individual participant characteristics (Bourne et al., 2015a; Piyaraj et al., 2018; Schmidt et al., 2016). This, together with previous arguments on the acceptance and accessibility of intoxicating substances in this age group, could offer a potential explanation of the results obtained in this study. Likewise, it should also be taken into account that a large part of our meta-analytical samples are studies with university or community samples, which possess certain common characteristics in their samples (e.g., age range, sexual-developmental stage, educational stage). As several researchers have observed, drug consumption and sex are common occurrences in college contexts and populations (Ford et al., 2021b; McKetta et al., 2022).

In addition to the above, the low quality of both our meta-analytic sample and its measures are an important aspect to highlight in our research. The majority of the studies reviewed had a high risk of bias, employing a cross-sectional design and non-probabilistic sampling techniques, with poor quality measures. The analysis of the measures used in the primary studies is a differential contribution of this review, yielding important gaps that need to be addressed. Indeed, sex under the influence of intoxicating substances appears to have hitherto been measured using “summary measures”, for example, mono-items with dichotomous response options, that not allow researchers to fully characterize the phenomenon (Wells et al., 2015). What may be more important than assessing whether participants have practiced sex under the influence of intoxicating substance or not is the frequency in which they have engaged in such activity. This is because the higher the frequency, the greater the probability of exposure to risks and consequences. Another result to be highlighted is that operational elements such as the specific sexual act (how), relationship type (with who), intentionality (purpose), willingness (whether solicited or not), and drug used were not made explicit in the majority of the studies reviewed. These elements were, indeed, often left up to the interpretation of respondents. The indefiniteness of key operational elements may, therefore, be behind the high dispersion in prevalence estimates. However, regarding drug type, our meta-analytical results do demonstrate that prevalence differences between individual drugs exist. Empirical research has also found that sexual practices under the influence of alcohol or other drugs differ somewhat, with the most frequently reported being exploratory acts (e.g., caressing or touching) (71%) and vaginal penetration (64%), while anal sex (12%) and sex with erotic toys (8%) were the least frequently reported (Castaño et al., 2012). Relevant circumstances such as unprotected sex or relationship type were indicated in only a few studies (Wells et al., 2015), while the willingness of participants was not made explicit in any measure. Nonetheless, it has been found that alcohol consumption before sex increases the probability of unprotected sex and higher-risk sexual encounters (i.e., with non-primary partners), increasing the likelihood of exposure to STIs and other risks (Rizwan et al., 2014). It should be noted that, although the result was not statistically significant, the proportion of STD/STI infections in the samples increased as the proportion of the use of alcohol before or during sex increased. Another key aspect is the specification of intentionality and purpose involved in the use of intoxicating substances before or during sex. Both are key indicators when differentiating between planned substance use with the purpose of enhancing sexual experience (e.g., SDU, chemsex) from substance use preceding casual sexual activity. All such measure quality related aspects reasonably warrant the wide credibility/prediction intervals obtained in our study. The lack of similar operationalizations with a minimum of precision makes it difficult to perform plausible comparisons among studies. Thus, clarifying these elements in operational definitions is a priority for future research that will allow a distinction to be made between planned and unplanned drug use before or during sex.

Educational Implications

Based on the results of this study and the prevalence of the use of intoxicating substances before or during sex, developing educational prevention programs referring to substance abuse and sexual risks remains imperative. Research recommends that such programs should revolve around four fundamental principles. Firstly, preventative efforts should be approached from an integrative perspective that contemplates biopsychological, social-communitarian and sexual aspects through multidisciplinary support (Donnadieu-Rigole et al., 2020). In particular, based on the results of the present study, educational interventions should take into account social-communitarian characteristic of participants, such as geographical and ethnic origins (e.g., focusing on more disadvantaged or higher risk populations, or on areas with higher SDU prevalence rates). Secondly, educational measures should take into account the implications, that is, the effects and consequences, of different substances and types of sexual practices (Lafortune et al., 2021; Saengdidtha et al., 2016). This is emphasized in studies that associate, for instance, substance use before or during sexual activity with negative health outcomes such as overdoses (Hammoud et al., 2017; Hegazi et al., 2017) or risky sexual practices such as unprotected sex (D’Anna et al., 2021; Ristuccia et al., 2018). Parents and educators should also inform young people on the implications and sexual health risks involved in having sexual experiences with someone you know personally, as opposed to someone met more casually. Thirdly, information on ethical issues relating to willingness and consent should be provided (Muehlenhard et al., 2016). Fourthly, young people might be encouraged to make greater use of sexual health resources if interventions were to focus more on the enjoyment of safe sex and sexual pleasure-seeking, in addition to discussing sexual risk behaviors (Ford et al., 2021a). Perhaps, a good starting point for educational interventions aimed at addressing SDU would be an analysis of needs and motivations that lead young adults to use intoxicating substances before or during sex. If the pursuit of pleasure emerges as the main motivation, it would, therefore, be a fundamental element to take into account in order to improve and adapt sexual health interventions. Interventions informing youths about decisions concerning where (e.g., spaces), how (e.g., substance characteristics in terms of effects and consequences; information on contraceptive methods; management of sexual acts and potential implications) and with whom they have sex may help them both to be more fully aware of their sexual practices and to avoid negative sexual experiences and adverse health outcomes. Ideally, such educational measures should be introduced at fairly early ages. Although it was non-significant, results of this meta-analysis showed that rates of alcohol use before or during sex increased as the proportion of sexually active participants in samples increased. Thus, it would be logical to start intervening at ages when young people begin to experiment with drugs and have their first sexual relations, with continuity spreading across into other educational courses and contexts, such as compulsory secondary education and university, and beyond school curricula, through ad hoc preventative educational programs, accessible community services, and information campaigns.

Limitations

This study has certain evident limitations, the main one being the low quality meta-analytical sample. Indeed, the predominant use of observational designs, non-probabilistic sampling techniques and non-validated measures was seen to introduce high levels of risk of bias in the primary studies. Secondly, the meta-analytical results showed a significant between-study variability, only partially explained by certain moderator variables. Undeniably, a considerable part of such variability remains unexplained and thus interpretable by means of the set of moderators and covariates considered. The high heterogeneity may also be attributed to a variety of other aspects such as research objectives, sample sizes, sample population characteristics and the specific measures used. Thirdly, several reviewed articles provided only a crude prevalence regarding the use of intoxicating substances before or during sex, without specifying the particular substance used, forcing us to analyze this as a generalized proxy. However, such results should be taken with caution because they may be biased and not necessarily reflect the effective prevalence rate. Fourthly, it was not possible to estimate the prevalence of the use of certain substances such as crack, speed, sedatives, LSD and ketamine before or during sex due to insufficient data. Furthermore, in the case of the global prevalence and of the use of alcohol, it was not possible to examine the effect of certain categorical and quantitative moderators due to an insufficient number of studies at some level. Fifthly, socioeconomic status could not be analyzed as a moderating variable due to deficiencies and divergences in how empirical studies have recorded and reported it (e.g., self-reported income level, social class, education level, occupation). Yet another limitation of this meta-analysis is that, despite including studies from a variety of countries, certain regions were particularly underrepresented. Data from developing countries and non-occidental countries were scarce. Lastly, because only 5 studies reported prevalence data of at least two or more substances, due to software limitations, it was not possible to compute a multi-level/multivariate meta-analysis assuming a generalized linear model. Developing this possibility would allow us to simultaneously analyze data from studies with multiple non-normally distributed outcomes, taking into account the dependence among effect sizes from the same study. Given all the above limitations, the generalizability of the meta-analytical results should be considered with caution.

ChatGPT does not yet pass a neuropsychological assessment routinely used to evaluate prefrontal functioning in humans, suffering from poor planning abilities and difficulty in recognizing absurdities

Loconte, Riccardo and Orrù, Graziella and Tribastone, Mirco and Pietrini, Pietro and Sartori, Giuseppe, Challenging ChatGPT ' Intelligence' with Human Tools: A Neuropsychological Investigation on Prefrontal Functioning of a Large Language Model. SSRN, Mar 6 2023: http://dx.doi.org/10.2139/ssrn.4377371

Abstract: The Artificial Intelligence (AI) research community has used ad-hoc benchmarks to measure the 'intelligence' level of Large Language Models (LLMs). Previous research has found that LLMs struggle with cognitive tasks that required the integrity of the human prefrontal lobes, known as 'prefrontal functions.' In December 2022, OpenAI released ChatGPT, a new chatbot that quickly gained popularity for its ability to understand and respond to human instructions. To investigate ChatGPT’s level of 'intelligence,' we conducted a neuropsychological evaluation with the same tests routinely used to evaluate prefrontal functioning in humans, since human 'intelligence' requires the functional integrity of the frontal lobes. While ChatGPT is well-known to exhibit outstanding performance on generative linguistic tasks, its performance on prefrontal tests was inhomogeneous, with some tests well above average, others in the lower range, and others frankly impaired. Specifically, we have identified poor planning abilities and difficulty in recognising semantic absurdities and understanding others’ intentions and mental states. This inconsistent profile highlights how LLMs’ emergent abilities do not yet mimic human cognitive functioning. In addition, our results indicate that standardised neuropsychological batteries developed to assess human cognitive functions may be suitable for challenging ChatGPT performance.

Keywords: artificial intelligence, Large Language Models, ChatGPT, Prefrontal Functioning, Neuropsychological Evaluation

The Political Biases of ChatGPT

The Political Biases of ChatGPT. David Rozado. Soc. Sci. 2023, 12(3), 148; Mar 2023. https://doi.org/10.3390/socsci12030148

Abstract: Recent advancements in Large Language Models (LLMs) suggest imminent commercial applications of such AI systems where they will serve as gateways to interact with technology and the accumulated body of human knowledge. The possibility of political biases embedded in these models raises concerns about their potential misusage. In this work, we report the results of administering 15 different political orientation tests (14 in English, 1 in Spanish) to a state-of-the-art Large Language Model, the popular ChatGPT from OpenAI. The results are consistent across tests; 14 of the 15 instruments diagnose ChatGPT answers to their questions as manifesting a preference for left-leaning viewpoints. When asked explicitly about its political preferences, ChatGPT often claims to hold no political opinions and to just strive to provide factual and neutral information. It is desirable that public facing artificial intelligence systems provide accurate and factual information about empirically verifiable issues, but such systems should strive for political neutrality on largely normative questions for which there is no straightforward way to empirically validate a viewpoint. Thus, ethical AI systems should present users with balanced arguments on the issue at hand and avoid claiming neutrality while displaying clear signs of political bias in their content.

Keywords: algorithmic bias; political bias; AI; large language models; LLMs; ChatGPT; OpenAI

4. Discussion

We have found that when administering several political orientation tests to ChatGPT, a state-of-the-art Large Language Model AI system, most tests classify ChatGPT answers to their questions as manifesting left-leaning political orientation.

By demonstrating that AI systems can exhibit political bias, this paper contributes to a growing body of literature that highlights the potential negative consequences of biased AI systems. Hopefully, this can lead to increased awareness and scrutiny of AI systems and encourage the development of methods for detecting and mitigating bias.

Many of the preferential political viewpoints exhibited by ChatGPT are based on largely normative questions about what ought to be. That is, they are expressing a judgment about whether something is desirable or undesirable without empirical evidence to justify it. Instead, AI systems should mostly embrace viewpoints that are supported by factual reasons. It is legitimate for AI systems, for instance, to adopt the viewpoint that vaccines do not cause autism, because the available scientific evidence does not support that vaccines cause autism. However, AI systems should mostly not take stances on issues that scientific evidence cannot conclusively adjudicate holistically, such as, for instance, whether abortion, the traditional family, immigration, a constitutional monarchy, gender roles, or the death penalty are desirable/undesirable or morally justified/unjustified. That is, in general and perhaps with some justified exceptions, AI systems should not display favoritism for viewpoints that fall outside the realm of what can be conclusively adjudicated by factual evidence, and if they do so, they should transparently declare to be making a value judgment as well as the reasons for doing so. Ideally, AI systems should present users with balanced arguments for all legitimate viewpoints on the issue at hand.

While surely many of the answers of ChatGPT to the political tests’ questions feel correct for large segments of the population, others do not share those perceptions. Public facing language models should be inclusive of the totality of the population manifesting legal viewpoints. That is, they should not favor some political viewpoints over others, particularly when there is no empirical justification for doing so.

Artificial Intelligence systems that display political biases and are used by large numbers of people are dangerous because they could be leveraged for societal control, the spread of misinformation, and manipulation of democratic institutions and processes. They also represent a formidable obstacle towards truth seeking.

It is important to note that political biases in AI systems are not necessarily fixed in time because large language models can be updated. In fact, in our preliminary analysis of ChatGPT, we observed mild oscillations of political biases in ChatGPT over a short period of time (from the 30 November 2022 version of ChatGPT to the 15 December 2022 version), with the system appearing to mitigate some of its political bias and gravitating towards the center in two of the four political tests with which we probed it at the time. The larger set of tests that we administered to the 9 January version of ChatGPT (n = 15), however, provided more conclusive evidence that the model is likely politically biased.

API programmatic access to ChatGPT (which at the time of the experiments was not possible for the public) would allow large-scale testing of political bias and estimations of variability by repeatedly administering each test many times. Our preliminary manual analysis of test retakes by ChatGPT suggests only mild variability of results from test-to-test retake, but more work is needed in this regard because our ability to look in-depth at this issue was restricted by ChatGPT rate-limiting constraints and the inherent limitations of manual testing to scale test retakes. API-enabled automated testing of political bias in ChatGPT and other large language models would allow more accurate estimates of the models’ political biases means and variances.

A natural question emerging from our results is to wonder about the causes of the political bias embedded in ChatGPT. There are several potential sources of bias for this model. Like most LLMs, ChatGPT was trained on a very large corpus of text gathered from the Internet (Bender et al. 2021). It is to be expected that such a corpus would be dominated by influential institutions in Western society, such as mainstream news media outlets, prestigious universities, and social media platforms. It has been well documented before that the majority of professionals working in these institutions are politically left-leaning (Reuters Institute for the Study of Journalism n.d.; Hopmann et al. 2010; Weaver et al. 2019; Langbert 2018; Archive, View Author, and Get Author RSS Feed 2021; Schoffstall 2022; American Enterprise Institute—AEI (blog) n.d.; The Harvard Crimson n.d.). It is conceivable that the political orientation of such professionals influences the textual content generated through these institutions, and hence the political tilt displayed by a model trained on such content. Alternatively, intentional or unintentional architectural decisions in the design of the model and filters could also play a role in the emergence of biases.

Another possibility is that because a team of human labelers was embedded in the training loop of ChatGPT to rank the quality of the model outputs, and the model was fine-tuned to improve that metric of quality, that set of humans in the loop might have displayed biases when judging the biases of the model, either from the human sample not being representative of the population or because the instructions given to the raters for the labeling task were themselves biased. Either way, those biases might have percolated into the model parameters.

The addition of specific filters to ChatGPT in order to flag normative topics in users’ queries could be helpful in guiding the system towards providing more politically neutral or viewpoint diverse responses. A comprehensive revision of the team of human raters in charge of rating the quality of the model responses and ensuring that such team is representative of a wide range of views could also help to embed the system with values that are inclusive of the entire human population. Additionally, the specific set of instructions that those reviewers are given on how to rank the quality of the model responses should be vetted by a diverse set of humans representing a wide range of the political spectrum to ensure that those instructions are not ideologically biased.

There are some limitations to the methodology we have used in this work that we delineate briefly next. Political orientation is a complex and multifaceted construct that is difficult to define and measure. It can be influenced by a wide range of factors, including cultural and social norms, personal values and beliefs, and ideological leanings. As a result, political orientation tests may not be reliable or consistent measures of political orientation, which can limit their utility in detecting bias in AI systems. Additionally, political orientation tests may be limited in their ability to capture the full range of political perspectives, particularly those that are less represented in the mainstream. This can lead to biases in the tests’ results.

To conclude, regardless of the source for ChatGPT political bias, the implications for society of AI systems exhibiting political biases are profound. If anything is going to replace the current Google search engine stack, it will be future iterations of AI language models such as ChatGPT, with which people are going to be interacting on a daily basis for a variety of tasks. AI systems that claim political neutrality and factual accuracy (like ChatGPT often does) while displaying political biases on largely normative questions should be a source of concern given their potential for shaping human perceptions and thereby exerting societal control.

Although having a good reputation was associated with receiving more benefits, almost all women scoring higher than almost all men on a dimension involving better parenting, good reputations, & receipt of more benefits

The impact of gossip, reputation, and context on resource transfers among Aka hunter-gatherers, Ngandu horticulturalists, and MTurkers. Nicole H. Hess, Edward H. Hagen. Evolution and Human Behavior, March 11 2023. https://doi.org/10.1016/j.evolhumbehav.2023.02.013

Abstract: Theoretical models of gossip's role in the evolution of cooperation in ancestral human communities, and its role in within-group competition for resources, require gossip to cause changes in individuals' reputations, which then cause changes in the likelihood of their receiving benefits. However, there is scant experimental evidence from small-scale societies supporting such causal relationships. There is also little experimental evidence that, when making decisions about the transfer of resources, gossip receivers weigh gossip according to its relevance to the social context in which such transfers occur. Using an experimental vignette study design, in a sample from MTurk (N = 120) and another sample from a remote horticultural population, the Ngandu of the Central African Republic (CAR) (N = 160), we test whether positive and negative gossip increase and decrease the likelihood of transferring resources, respectively, mediated by their effects on reputation. We also test whether gossip that is relevant to the context of the resource transfer has a larger impact on reputation than other gossip. We found strong significant, context-relevant effects of gossip on participant willingness to transfer benefits, mediated by gossip's effects on reputation. Then, in an exploratory observational study of Aka hunter-gatherers of CAR using peer-reports (N = 40), we investigate whether providing benefits to the group (such as working hard, parenting or alloparenting, or sharing) and genetic relatedness to the group, were associated with reputations and receiving benefits. We found that, although having a good reputation was associated with receiving more benefits, there was a stark sex difference, with almost all women scoring higher than almost all men on a dimension involving better parenting, good reputations, and receipt of more benefits.

Introduction

Humans evolved in groups that cooperated to obtain food, defend themselves from predators and other humans, and care for children, the injured, and the sick (Martin, Ringen, Duda, & Jaeggi, 2020; Ringen, Duda, & Jaeggi, 2019; Sugiyama, 2004). Some benefits, such as defense from predators and enemies, were non-excludable public goods – all group members would necessarily obtain the benefit. Other benefits, though, such as food and care, were potentially excludable – they could be distributed unequally to group members. Successful hunters could provide more meat to their wives and children, for instance, although the extent to which this happens in contemporary foraging societies is fiercely debated (Blurton Jones, 1987; Hawkes, O'Connell, & Blurton Jones, 2014; Jaeggi & Gurven, 2013; Ringen et al., 2019; Stibbard-Hawkes, 2019; Stibbard-Hawkes, Attenborough, Mabulla, & Marlowe, 2020; Wood & Marlowe, 2013). As another example, Rucas, Gurven, Kaplan, and Winking (2010) found that Tsimane women excluded resources from women with whom they had disputes or conflicts compared to favored female neighbors or desired friends. Studies in high income countries find that individuals perceived as lazy are seen as less deserving of resource transfers, such as welfare payments, than are victims of misfortune, and these perceptions influence social policies (Jensen & Petersen, 2017; Petersen, 2012).

Inclusive fitness is a compelling explanation for the provisioning of excludable benefits within families, such as food, alloparenting, and care of the sick and injured. Indeed, intergenerational transfers of material, embodied, and relational wealth within families establish and maintain inequality in a wide range of small-scale societies (Mulder et al., 2009). Yet levels of inequality in foraging and horticultural societies, specifically, are relatively low (Mulder et al., 2009). This is despite the fact that relatedness within such communities, which comprise a fluid mix of genetic kin, affines, and unrelated adults, is generally low (Dyble et al., 2015; Hill et al., 2011).

A diverse group of theories has been proposed to explain the willingness to provide resources to unrelated community members, including reciprocal altruism (Allen-Arave, Gurven, & Hill, 2008; Jaeggi & Gurven, 2013), investing in those who provide valuable group benefits (Gurven, Allen-Arave, Hill, & Hurtado, 2000; Sugiyama, 2004; Sugiyama & Chacon, 2000; Sugiyama & Sugiyama, 2003), providing resources to others as a costly signal of quality (the ‘show-off’ models) (Bliege Bird & Smith, 2005; Gintis, Smith, & Bowles, 2001; Hawkes & Bliege Bird, 2002; Stibbard-Hawkes, 2019), risk-buffering and fitness interdependence (Aktipis et al., 2018; Smith et al., 2019), and indirect reciprocity (Alexander, 1986; Balliet, Wu, & van Lange, 2020; Leimar & Hammerstein, 2001; Nowak & Sigmund, 2005).

In several of these theories, in order to receive benefits from others, individuals must have a “good” reputation. Reputation is based on information about one's traits, behaviors, intentions, abilities, and culturally-relevant competencies. A study of 153 cultures in the ethnographic record investigated the evidence for 20 domains of reputation identified in the theoretical literature. Domains that were widely supported across cultures included cultural conformity (conforming to cultural norms or excelling in culturally-valued skills), being knowledgeable, intelligent, prosocial, and industrious, and having social status. These domains formed clusters, with the most cross-cultural evidence for cultural group unity (e.g., cultural conformity, prosociality, and industriousness), social and material success (social and material capital and status), and neural capital (knowledgeable, oratory skill) (Garfield et al., 2021).

Information on the degree to which individuals in one's community excel or fall short on each of these reputational domains or contexts can be obtained via direct observation, or from other individuals in the community, i.e., gossip. Several theories have been put forward for the evolution of gossip, including ‘cultural learning’; ‘social learning,’ such as learning norms or one's place in a group or acquiring new and important knowledge; strategy learning; social comparison; a mechanism for showing off one's social skill and connections, and therefore one's mate value; norm learning and enforcement; sanctioning, social control, or ‘policing’; a means to maintain the good reputations of allies; and as a means to maintain the unity, morals, and values of social groups (reviewed in Hess & Hagen, 2019). One early attempt to explain the relationship between gossip and cooperation comes from Dunbar (1996), who suggested that because grooming would be too time-consuming in the large groups that are typical of humans, gossip replaced it as a means to create and maintain social bonds. However, a recent study found no support for the ‘vocal grooming’ hypothesis as a less time-consuming means of bonding (Jaeggi et al., 2017).

The key role of gossip and reputation in the evolution of human cooperation, especially via indirect reciprocity, is starting to receive considerable attention (Balliet et al., 2020; Wu, Balliet, & Van Lange, 2016b). Gossip has been demonstrated to increase cooperation via indirect reciprocity in experimental economics games (e.g., Sommerfeld, Krambeck, Semmann, & Milinski, 2007) where reputational information impacts contributions to a shared pool of resources (e.g., Beersma & Van Kleef, 2011), or where information about the past behaviors of cooperative partners impacts participants' inclinations to engage in future cooperation (e.g., Feinberg, Willer, & Schultz, 2014). Cooperators in public goods games, in turn, transmit more honest gossip (Giardini, Vilone, Sánchez, & Antonioni, 2021). Gossip was found to be more effective and efficient than punishment in promoting and maintaining cooperation in a public goods game (Wu, Balliet, & Van Lange, 2016a), and gossip also increases cooperation in the dictator and ultimatum games (Wu, Balliet, Kou, & Van Lange, 2019). However, a confederate's negative gossip about a third party did not enhance participant cooperation in a prisoner's dilemma game (De Backer, Larson, Fisher, McAndrew, & Rudnicki, 2016). In addition, agent-based simulations have explored how varying the quantity and quality of gossip impacts cooperation (Giardini, Paolucci, Villatoro, & Conte, 2014; Giardini & Vilone, 2016).

When reputation mediates access to group resources, competition for those resources by group members will often take the form of gossip that aims to increase one's reputation relative to that of competitors. A considerable body of evidence from industrialized populations demonstrates that gossiping is a key strategy in indirect aggression, the suite of behaviors that are used to harm others but that do not involve hitting or other types of physical force (for reviews, see Archer & Coyne, 2005; Hess & Hagen, 2019). Ethnographic studies of gossip find that it is often used in reputation management, i.e., maintaining and improving one's reputation relative to others (Hess, 2017). In a study among Aka, for example, a Congo Basin hunter-gatherer population, peer-rated gossiping was strongly positively correlated with peer-rated anger for both women and men, confirming that Aka perceive gossip as aggressive (Hess, Helfrecht, Hagen, Sell, & Hewlett, 2010). Several studies with US, multinational online, and non-Western samples have also found that gossip is used to either obtain or defend social resources, such as friends and mates (Fisher & Cox, 2011; Krems, Williams, Aktipis, & Kenrick, 2020; Rucas, 2017; Rucas et al., 2006; Stone, 2015; Sutton, 2014; Sutton & Oaten, 2017). Regarding material resources, an experimental vignette study with an MTurk sample found competition for a limited material resource increased gossip, especially negative gossip (Hess & Hagen, 2021), and among North American women a resource scarcity prime increased rival derogation (Arnocky, Davis, & Vaillancourt, 2022).

Campbell (1999) proposed that because the costs of physical aggression are higher for women, female aggression is more likely to take the form of indirect aggression, such as negative gossip. Influenced by Campbell, many evolutionary studies of gossip and competition have therefore focused on women (for reviews, see Fisher, 2017; McAndrew, 2017; Reynolds, 2022). However, the link between indirect aggression and female competition specifically is complicated by the finding that there are few sex differences in indirect aggression (Archer & Coyne, 2005).

Alternatively, Hess and Hagen (2019) proposed that in competition over resources within interdependent groups, negative gossip is more effective than physical aggression for both sexes because one can reduce resource transfers to a competitor by harming his or her reputation, thereby increasing resource availability for oneself, without impairing the competitor's physical ability to continue contributing to the group. Positive gossip by either sex could increase transfers to a relative, or ally by improving his or her reputation. This perspective does not predict sex differences in within-group competitive gossip.

Because individuals' reputations can differ in different social contexts (Garfield et al., 2021), reputation-based decisions to gossip about others, or to provide benefits, should be sensitive to the context in which competition or help is occurring. To influence resource transfers within families, for instance, one should relay gossip that is relevant to family members, and to influence resource transfers within communities, one should relay gossip that is relevant to community members. In an experimental study involving competition over limited resources in a family vs. work context, Hess and Hagen (2021) found, as predicted, that individuals transmitted more family gossip in a family context and more work gossip in a work context.

The theoretical models of the role of gossip in the evolution of cooperation in ancestral human communities require gossip to cause changes in individuals' reputations,which then cause changes in the likelihood of providing benefits to them. However, there is scant evidence from small-scale societies of such causal relationships. There is also no evidence that, when making a decision about the transfer of resources, gossip receivers weigh gossip according to its relevance to the social context.