AWEI latest news and analysis

AI-compiled news and analysis across business, science, health and other topics, with publisher sources and links to the original reports.

Analysis · Technology

More HR Automation Did Not Reliably Signal Stronger AI Ethics

An automated HR workflow can have human reviewers without scoring highly on ethical safeguards. In a Nepalese survey, automation was associated with reported human oversight, but its relationship with reported ethical AI practices was not statistically significant after accounting for general AI adoption and analytics use. These were distinct measures of governance. Manoj Varghese and Umesh Jeet Sharma’s 2026 study surveyed 260 HR professionals, line managers and employees across several sectors in Nepal. Participants answered five-point questionnaires about technology use, governance and perceived decision effectiveness. Ethical practices covered issues including transparency, accountability and bias mitigation; oversight concerned human discretion and judgment over AI decisions. The researchers used separate regression models to examine which technology measures were associated with…

Read article: More HR Automation Did Not Reliably Signal Stronger AI Ethics
Share on Facebook: More HR Automation Did Not Reliably Signal Stronger AI Ethics

Analysis · Business

What a Livestream Commerce Checklist Can Leave Out

A payment problem mentioned once can still matter to a merchant. But in a 2023 study of China’s cross-border food livestream commerce, concepts appearing fewer than three times—or judged mutually contradictory—were discarded during analysis. That methodological choice matters when turning the resulting framework into a business checklist. Gaofeng Wang and colleagues used grounded theory to build an explanation from documents. Their material included policy texts, research reports, news, literature and government-recognized business cases. They coded two-thirds of the material into 48 concepts, then 17 categories and six broad areas covering platforms, livestream commerce, agriculture, demand, support services and the external environment. The useful shift is from evaluating a broadcast to examining the transaction around it. Their framework includes payment systems, logistics, staff…

Read article: What a Livestream Commerce Checklist Can Leave Out
Share on Facebook: What a Livestream Commerce Checklist Can Leave Out

Analysis · Business

AI Service Growth Depends on What Can Be Reused—and What Must Change

A system for managing underground mine traffic may also help map the mine. In a 2021 study, the manufacturer given the pseudonym Rockcorp described this additional capability of its AI application. The paper does not establish that the mapping service was deployed. But the example raises a useful growth question: how much of the next offering is already built? David Sjödin, Vinit Parida, Maximilian Palmié and Joakim Wincent explored six large, globally active Swedish industrial manufacturers and their partners. They compared interview themes with company reports, agreements and project documents, developing a framework for expanding AI services. This was an exploratory case study, with no experimental test of whether reuse improved profits. The interviews revealed a tension between customization and expansion. Customers needed services suited to their particular operations; providers…

Read article: AI Service Growth Depends on What Can Be Reused—and What Must Change
Share on Facebook: AI Service Growth Depends on What Can Be Reused—and What Must Change

Analysis · Finance

A Better Annual Forecast Score Can Hide Worse Winter Coverage

Which forecast range is better: one that contains 93% of outcomes, or one that contains 91%? If both promise 95%, the first looks closer to its target. But a 2024 methods study by Ricarda Duerst and Jonas Schöley shows why that annual comparison can miss an important weakness. Using weekly death records from 23 European countries, the researchers repeatedly fitted a forecasting model to five years of data and predicted the following two years. They used errors from earlier forecasts to build uncertainty ranges, allowing the spread and asymmetry of errors to change with the season. Separate validation periods tested those ranges against unseen observations. The conventional negative-binomial intervals achieved 93% coverage across the year, compared with 91% for the fitted empirical intervals. Neither reached the stated 95% target. Yet the conventional intervals contained only 85% of…

Read article: A Better Annual Forecast Score Can Hide Worse Winter Coverage
Share on Facebook: A Better Annual Forecast Score Can Hide Worse Winter Coverage

Analysis · Finance

How Easier Market Entry Could Push Established Firms to Improve

A new manufacturer can change the incentives of firms already operating. That is the mechanism explored by Helu Jiang, Yu Zheng and Lijun Zhu in a manuscript dated 2022, issued as a working paper in 2023: easier entry may support growth by making market leaders face closer competitors. The researchers built an economic model and calibrated it to Chinese manufacturing data from 2004–2007, using industrial census and enterprise survey records. Each model industry contains two competing firms and a potential entrant. Firms invest in productivity improvements, differ in growth potential, and face an administrative hurdle before entering. They then simulated an alternative economy with entry barriers estimated to resemble those of 1990. Annual productivity growth fell from 5.87% in the baseline model to 4.42%. The difference represents 24.7% of baseline growth. This is a model-based estimate…

Read article: How Easier Market Entry Could Push Established Firms to Improve
Share on Facebook: How Easier Market Entry Could Push Established Firms to Improve

Analysis · Finance

A Labor Income Signal That Explained Returns Better Than It Forecast Them

Companies with slower labor income growth had higher average stock returns in a Pakistan study. But the more revealing result came when researchers tried to turn historical relationships into forecasts: their model’s overall predictive tests fell short of statistical significance at the 5% level. Published in 2023, the study by Naveed Khan and colleagues examined 164 non-financial companies listed on the Pakistan Stock Exchange from July 2010 to June 2020. It combined monthly share prices with company accounting data to test an extension of the Fama-French three-factor model. The researchers sorted companies into eight portfolios using market capitalization, book-to-market ratios and labor income growth. They added a fourth factor: the return difference between portfolios with low and high labor income growth. This was their proxy for human capital, rather than a direct measurement of…

Read article: A Labor Income Signal That Explained Returns Better Than It Forecast Them
Share on Facebook: A Labor Income Signal That Explained Returns Better Than It Forecast Them

Analysis · Education

A Reputation for Credibility Is Not a Test of Truth

A restaurant blog can seem credible without its claims ever being checked. That gap is central to reading a 2022 study by Zehra Ece Serman and Julian Sims: reputation and perceived trustworthiness were associated with perceived credibility, but the researchers did not establish whether the posts were accurate. The study collected 449 online survey responses from UK residents between December 2020 and February 2021. Recruitment targeted university students through social media and participants’ contacts. Respondents answered questions about hospitality blogs using five-point scales. The researchers used structural equation modelling to examine relationships among groups of answers representing credibility, expertise, reputation and other perceptions. The unexpected result was that perceived expertise had no statistically significant association with credibility in the fitted model.…

Read article: A Reputation for Credibility Is Not a Test of Truth
Share on Facebook: A Reputation for Credibility Is Not a Test of Truth

Analysis · Science

How Young Adults Became a Reference for Vaccination Readiness

To estimate how many older Brazilians vaccination protected, researchers watched what happened to people aged 20–29. This comparison gave Leonardo Souto Ferreira and colleagues a way to separate a receding epidemic from changes associated with vaccination in their study published online in 2022. Falling hospitalisations alone cannot reveal a campaign’s contribution: the epidemic itself may be easing. The team fitted a Bayesian statistical model to national surveillance and vaccination records, separately for Brazilian states and older age groups. Younger adults supplied a reference trajectory; increasing vaccine coverage helped explain departures from that trajectory among adults aged 60 and over. The crucial assumption was that, without vaccination, outcomes in older and younger groups would maintain a stable relationship. Coverage entered the model 14 days after second doses.…

Read article: How Young Adults Became a Reference for Vaccination Readiness
Share on Facebook: How Young Adults Became a Reference for Vaccination Readiness

Analysis · Health

Belonging and Health: What an Immigrant Support Study Found

Feeling part of a neighborhood did not consistently line up with better health in a study of immigrants in Málaga, Spain. Stronger belonging was associated with greater life satisfaction, but also with more reported physical illness. That unexpected combination complicates the idea that community connection automatically means healthier lives. In their 2019 study, Isabel Hombrados-Mendieta and colleagues surveyed 1,131 immigrants aged 18–70. Trained interviewers used random routes through city districts to recruit participants at places including businesses, immigrant associations and social service centers. Questionnaires were translated for participants who did not speak Spanish. The researchers distinguished support from family, friends born in Spain and fellow immigrant friends. They asked about emotional support, practical help and useful information, measuring both frequency and…

Read article: Belonging and Health: What an Immigrant Support Study Found
Share on Facebook: Belonging and Health: What an Immigrant Support Study Found

Analysis · Health

Why 110 COVID-19 Studies Still Left Major Evidence Gaps

A review can gather 110 studies and still leave most treatment questions uncertain. In their 2020 review of COVID-19 medicines, Min Seo Kim and colleagues rated 71% of the assessed evidence as low or very low certainty. The volume of research concealed substantial gaps in what it could establish. The team searched databases, trial registries and preprint sources through August 24, 2020. They included 40 randomized trials and 70 observational studies. Their efficacy analysis concerned adults hospitalized with COVID-19, separating patients initially outside intensive care from critically ill patients. Those settings mattered because illness severity could affect treatment responses. They used network meta-analysis, which combines direct treatment comparisons with indirect comparisons through shared alternatives, such as standard care. This allowed them to compare many regimens without…

Read article: Why 110 COVID-19 Studies Still Left Major Evidence Gaps
Share on Facebook: Why 110 COVID-19 Studies Still Left Major Evidence Gaps

Loading article…