Low AI Research Articles

Students’ prompt patterns and its effects in AI-assisted academic writing: Focusing on students’ level of AI literacy

The increase of generative artificial intelligence (GenAI) in education has emphasized prompt literacy and design, influencing the experience of GenAI-assisted learning. However, there is limited understanding regarding the various prompt patterns that emerge during student-GenAI interaction (SAI) within the context of navigating learning tasks at different levels of AI Literacy or how these patterns are associated with task performance. This study analyzed 19 university students’ prompt patterns when interacting with GenAI on academic writing tasks, grouping students by AI literacy level and examining how prompt patterns relate to performance. This study performed content analysis of student-GenAI chat histories and categorical data analysis of student interviews. Pattern differences were then visualized using Gephi 0.10.1. Distinctive prompt patterns were identified: high AI literacy students exhibited descriptive, context-based prompts in collaborative interactions, while low AI literacy students demonstrated general prompts in a student-directed approach. Additionally, student essays were evaluated by five experts to inform how prompt patterns relate to task performance and the Wilcoxon signed-rank test was applied. The two groups demonstrated significant differences across categories of writing performance (content, structure, and expression). This study provides implications for designing and implementing GenAI and student-centered AI-assisted instruction.

The Real Environment Impact of AI: Unveiling the Ecological Footprint of Artificial Intelligence

Global environmental pollution has a devastating influence on the planet's population and jeopardizes humanity's future. The construction business is a major producer of waste and hazardous emissions into the atmosphere. It is vital to discover measures to reduce the damage done to nature. Currently, artificial intelligence technologies are one of the most promising approaches to helping the environment. This research investigates the use of green AI algorithms for measuring greenhouse gas (GHG) emissions in the context of ecological footprint assessment. Green AI algorithms prioritize sustainability and seek to lower AI systems' carbon footprints while monitoring GHG emissions data. These algorithms use environmentally conscious machine learning techniques to improve resource allocation, encourage energy-efficient model topologies, and prioritize renewable energy sources for AI model training. Carbon-aware optimization approaches are used to reduce the environmental impact of AI computations, resulting in a greener future. The incorporation of green algorithms into AI systems identifies the potential for emission reduction and energy efficiency, promoting environmentally beneficial behaviours across industries. The use of green algorithms allows for a full analysis of GHG emissions and ecological footprints, permitting a symbiotic interaction between technology and the environment for sustainable growth.

The Beneficial Dietary Effect of Dried Olive Pulp on Some Nutritional Characteristics of Eggs Produced by Mid- and Late-Laying Hens.

This research evaluated the impact of incorporating dried olive pulp (OP) into the feed of laying hens on the fatty acid profile, cholesterol, triglyceride, total phenolic, oleuropein and hydroxytyrosol content, and health lipid indices of eggs produced by mid- (39 weeks) and late-laying (59 weeks) birds. Over a 36-week trial, 300 eggs from 180 Isa-Brown hens, assigned to three dietary groups with different OP levels (CON, OP4 and OP6), were analyzed. OP reduced egg cholesterol, with significant effects in late-age eggs (p < 0.05). In mid-age hens, the OP6 eggs had higher total phenolics than the controls (p < 0.05) and more PUFAs than the other groups (p < 0.05). The concentration of total phenolics, cholesterol, n3 PUFAs and % fat increased with hen age (p < 0.05), while triglycerides and oleuropein decreased (p < 0.05). With increasing hen age, the SFAs in the OP eggs decreased (p < 0.05) and the MUFAs increased (p < 0.05). Eggs from older hens had higher nutritional value, as indicated by the lower n6/n3 PUFA ratio, lower AI and TI indices, and higher h/H ratio (p < 0.05). Overall, dietary OP supplementation improved the nutritional quality of eggs, suggesting potential health benefits. Our results also highlighted eggs from older hens as a valuable source of high-quality fats.

Shaping the future: perspectives on the Integration of Artificial Intelligence in health profession education: a multi-country survey

BackgroundArtificial intelligence (AI) is transforming health profession education (HPE) through personalized learning technologies. HPE students must also learn about AI to understand its impact on healthcare delivery. We examined HPE students’ AI-related knowledge and attitudes, and perceived challenges in integrating AI in HPE.MethodsThis cross-sectional included medical, nursing, physiotherapy, and clinical nutrition students from four public universities in Jordan, the Kingdom of Saudi Arabia (KSA), the United Arab Emirates (UAE), and Egypt. Data were collected between February and October 2023 via an online survey that covered five main domains: benefits of AI in healthcare, negative impact on patient trust, negative impact on the future of healthcare professionals, inclusion of AI in HPE curricula, and challenges hindering integration of AI in HPE.ResultsOf 642 participants, 66.4% reported low AI knowledge levels. The UAE had the largest proportion of students with low knowledge (72.7%). The majority (54.4%) of participants had learned about AI outside their curriculum, mainly through social media (66%). Overall, 51.2% expressed positive attitudes toward AI, with Egypt showing the largest proportion of positive attitudes (59.1%). Although most participants viewed AI in healthcare positively (91%), significant variations were observed in other domains. The majority (77.6%) supported integrating AI in HPE, especially in Egypt (82.3%). A perceived negative impact of AI on patient trust was expressed by 43.5% of participants, particularly in Egypt (54.7%). Only 18.1% of participants were concerned about the impact of AI on future healthcare professionals, with the largest proportion from Egypt (33.0%). Some participants (34.4%) perceived AI integration as challenging, notably in the UAE (47.6%). Common barriers included lack of expert training (53%), awareness (50%), and interest in AI (41%).ConclusionThis study clarified key considerations when integrating AI in HPE. Enhancing students’ awareness and fostering innovation in an AI-driven medical landscape are crucial for effectively incorporating AI in HPE curricula.

A comprehensive foodomics analysis of rambutan seed oils: Focusing on the physicochemical parameters, lipid concomitants and lipid profiles

A foodomics approach was employed to systematically characterize and compare the quality parameters, antioxidant activity, minor-components, fatty acid composition, and lipid profiles of the seed oils from the three most popular rambutan varieties in China. The total lipid content ranged from 23.40 to 25.77 g/100 g. The fatty acids 9cC18:1 (39.84%–40.92%) and C20:0 (28.45%–30.23%) were identified as the dominant ones, which are uncommon among higher plants. All oil samples exhibited low AI and TI values. BR-7 exhibited the highest levels of squalene (21.48 mg/kg), cholesterol (144.43 mg/kg), and tocopherol (17.42 mg/kg), and the lowest levels of polyphenols (24.21 mg GAE/kg). Additionally, a total of 807 lipid species were identified, with TAG, DGTS, and PE being the predominant ones. Multivariate statistical analyses revealed significant variations in lipid profiles among the varieties, particularly in glycerophospholipids and sphingolipids. Fifty-seven distinct lipids were identified as potential markers for distinguishing between rambutan varieties. Furthermore, a hypothetical scenario was developed by linking relevant lipid metabolism pathways. These findings establish a theoretical framework for comprehending rambutan seed oil in depth and unlocking its high-value potential.

Assessing human risks and benefits associated with mercury contamination versus fatty acids and lipid quality indices in Anguilla anguilla L., Perca fluviatilis L. and Rutilus rutilus L. from the Olsztyn Lake District, Poland

Abstract The concentration of mercury, fatty acids and lipid quality indices were determined in three species of fish (European eel, Anguilla anguilla Linnaeus, 1758, Eurasian perch, Perca fluviatilis Linnaeus, 1758 and roach, Rutilus rutilus Linnaeus, 1758). The fish were caught from a reservoir in the Warmia and Mazury region located near Olsztyn (north-eastern Poland). Mercury content was determined by atomic absorption thermal decomposition using Milestone DMA-80 (Italy), whereas fatty acids were measured by gas chromatography using a 7890A Agilent Technologies chromatograph. The total mercury concentration had the following sequence: European perch (0.297 mg/kg) &gt; European eel (0.160 mg/kg) ≈ roach (0.100 mg/kg) (P&lt;0.05). A positive correlation was found between the mercury content and size (the body weight and total length) of these fish (P&lt;0.0001). Based on the THQ, HI (&lt;1.0) and BRQ (1.063), it was concluded that the fish does not pose a health risk and is fit for human consumption. Similarly, low AI and TI values (&lt;1) are beneficial to human health related to reducing the risk of ischemic heart disease. A low OFA index of C12+C14+C18 and a high index of C18+UFA also indicate that these fish are safe from a nutritional point of view. The fatty acid indices HH (32.10–74.30), NVI (0.83–2.14), DI(18) (65.50–89.28), DI(16) (18.45–32.96), UI (131.93–278.64) and PI (88.60–314.90) were also studied. However, higher PUFA/SFA and de minimis EPA+DHA/Hg ratios compared to the recommended ratios suggest further research in this area is needed.

AI as Co-Leader: Effects of Human Consideration and AI Structure Behaviors on Leadership Effectiveness and Neural Activation

With rapid technological advancements, AI has emerged as a potential performer of some leadership functions, supplementing or even substituting certain human leadership. The collaborative leadership dynamics between humans and AI remain underexplored. This study investigated the effects of human consideration (high versus low) and AI structure (high versus low) behaviors on task performance, subjective experiences, and neural activation in a human–AI co-lead scenario. A total of 80 participants engaged in quality inspection tasks that were co-led by human and AI leaders, with 20 participants randomly allocated to each behavior combination. The results revealed that high AI structure was associated with increased job-related satisfaction, employee engagement, and leader trust, while low AI structure slightly facilitated task completion speed (about 6.80%). High human consideration increased the trust in and satisfaction with the human leader. Overall, the combination of high consideration in the human leader and high structure in the AI leader enhanced leadership effectiveness and diversification of work methods (with 30% of participants choosing inspection methods other than AI suggestions). Low-structure AI leader behavior led to increased neural activation during the human-leading process in the prefrontal cortex, which was associated with increased attention, motivation, and semantic processing.

Scenario-Based Approach to AI’s Agency to Perform Human-Specific Tasks

Abstract The development of service robots and chatbots has changed the way companies interact with their customers. Several automated tasks have been taken over by AI, such as communication with customers, processing of orders and also other automated tasks. To measure the impact AI will have on our society, we have analyzed AI’s agency by testing different scenarios of AI’s ability to perform human-specific tasks such as having own political opinions, religious beliefs and other situations which go beyond the simple execution of tasks. The results show that consumers perceive a low probability that AI has such thoughts. However, AI with stronger anthropomorphic characteristics is more likely to have agency to perform these types of tasks. According to our study, a human-like appearance enhances the perception that AI can have political views, religious beliefs and an own identity. Moreover, people consider that AI with anthropomorphic characteristics should be responsible for their mistakes and punished if they do something wrong. In spite of this, AI, independent of its anthropomorphic characteristics, is not believed to have the ability to solve mankind problems such as reducing pollution, stabilizing the world economy or to ensure world peace. Our research also shows that men are more inclined to attribute agency to AI compared to women, which perceive a lower AI agency.

Can I Trust What You Say about the Smart Speaker? Effects of AI Anxiety and Biased Assimilation on Blog Comment Credibility

ABSTRACT Advancements in Internet and communication technologies have facilitated the proliferation of product reviews and commentary pertaining to smart speakers. This study investigated the impacts of uncivil comments on blog comment credibility. Results show that the civil opinion commentary appears to be more credible than uncivil commentary. The interaction between civil opinion and low AI anxiety generates more blog review credibility than the interaction between uncivil opinion and low AI anxiety. Finally, we identify moderating effects of AI anxiety and provide deeper insights into how the interaction of comment civility and AI anxiety shape the credibility assessments of blog reviews.

Metal Content, Fatty Acid and Vitamins in Commercially Available Canned Fish on the Bulgarian Market: Benefit-Risk Ratio Intake.

Today, the distribution and consumption of canned products have become widespread due to the convenience of using processed fish products. This study aims to evaluate elements of toxicological concern and essential elements (Cd, Al, Cu, Cr, Mn, Fe, Pb, Ni, and Zn), the fatty acid composition, and the fat-soluble vitamin and antioxidant pigment contents of various canned fish products purchased on the Bulgarian market. The estimated weekly intake and human health benefit-risk ratio based on metal elements and n-3 LC-PUFA contents in canned fish products were assessed. The contents of the analyzed elements in the canned samples were found to be below the limits set by various health organizations. Al was detected in only one sample. The profile of fatty acids showed that the canned fish had high PUFA/SFA ratios, EPA + DHA contents and low SFA, AI and TI values. The analyzed canned fish contained significant contents of fat-soluble vitamins. The Atlantic bonito in a jar sample was characterized by a high percentage of vitamin A (26.7% RDI) and vitamin D3 (142% RDI) per 100 g portion. The calculated EWI value shows that the consumption of canned fish products did not present any risk. The benefit-risk ratio indicates that the canned fish species are safe for human consumption, except for one sample regarding Cr.

Dependence of annual asymmetry in &lt;i&gt;NmF2&lt;/i&gt; on local time

Based on the global empirical model of the F2 layer critical frequency median (Satellite and Digisonde Data Model of the F2 layer, SDMF2), an analysis was made of the properties of diurnal variations in the annual asymmetry in the concentration of the F2 layer maximum NmF2 at different values of the solar activity index F. The AI index, which characterizes the relative difference in NmF2 averaged over all longitudes and latitudes between January and July at a given local time, was used as a parameter of this asymmetry. It was found that the diurnal variations of the AI index are dominated by a semidiurnal mode with maxima in the daytime and at night. The daytime maximum of the AI index is almost independent of the level of solar activity. The nighttime AI maximum decreases with increasing solar activity. For low solar activity, the daytime and nighttime AI maxima almost coincide in amplitude when AI = 16—17%. The difference in the solar radio flux between January and July due to the ellipticity of the Earth’s orbit relative to the Sun makes a significant contribution to the AI index at all hours of the day. On average, it is 3—4% and can reach 5% with low solar activity at night. The difference in the AI index for low and high activity according to the IRI model (with URSI and, especially, CCIR coefficients) is overestimated relative to the SDMF2 model at almost all hours of the day, apparently due to the limited number of experimental data when obtaining the CCIR and URSI coefficients especially over the oceans

Accelerated Low Power AI System for Indian Sign Language Recognition

Accelerated Low Power AI System for Indian Sign Language Recognition

Customers’ intention to compliment and complain via AI-enabled platforms: A self-disclosure perspective

Building upon the communication privacy management theory, the research reveals the effect of self-disclosure on the identified mechanisms of perceived emotional value, performance expectancy, and privacy concerns, which in turn, influence customers' intention to compliment and complain via AI-enabled platforms. Findings from two quasi-experiments with 439 valid responses from U.S. customers suggest that customers are more likely to express their feelings when low self-disclosure AI technology is presented. The results suggest a prominent role of privacy concerns in mediating the effect of self-disclosure on customers’ intention to compliment and complain. The effects of self-disclosure also channel through perceived emotional value and performance expectancy when customers want to leave a compliment. The moderating effect of reward timing was examined. Similarities and differences between customers’ intentions to compliment or complain using AI-enabled platforms are discussed to provide theoretical and practical implications.

Overview of trials on artificial intelligence algorithms in breast cancer screening – A roadmap for international evaluation and implementation

Accumulating evidence from retrospective studies demonstrate at least non-inferior performance when using AI algorithms with different strategies versus double-reading in mammography screening. In addition, AI algorithms for mammography screening can reduce work load by moving to single human reading. Prospective trials are essential to avoid unintended adverse consequences before incorporation of AI algorithms into UK’s National Health Service (NHS) Breast Screening Programme (BSP). A stakeholders’ meeting was organized in Newnham College, Cambridge, UK to undertake a review of the current evidence to enable consensus discussion on next steps required before implementation into a screening programme.It was concluded that a multicentre multivendor testing platform study with opt-out consent is preferred. AI thresholds from different vendors should be determined while maintaining non-inferior screening performance results, particularly ensuring recall rates are not increased. Automatic recall of cases using an agreed high sensitivity AI score versus automatic rule out with a low AI score set at a high sensitivity could be used. A human reader should still be involved in decision making with AI-only recalls requiring human arbitration. Standalone AI algorithms used without prompting maintain unbiased screening reading performance, but reading with prompts should be tested prospectively and ideally provided for arbitration.

More Than Just a Memory

This article examines American Artist’s installation “Sandy Speaks,” a chatbot inspired by Sandra Bland’s media activism, as an analytical connective point to pathways of Black technological critique and histories of Black digitality and care. First connecting the work to its predigital antecedent The Negro Green Book, the article then argues that the typical aspirations of chatbot to approximate the human is disavowed in “Sandy Speaks,” enacting a Black technological critique of the human itself. Moreover, departing from celebratory discourses of Black technological innovation, the chatbot's low AI instantiates what the author calls a politics of technological refusal—a praxis of deliberate technological limitation as critique. This article asks what might happen when we seek potentialities of Black praxis in the slow, broken, old, technological forms, not as remedy, but as theory, critique, and an undoing of the recuperation of technological innovation as most legible mode of recognition.

A multicenter evaluation on the impact of non-therapeutic transfer in rural trauma

IntroductionTrauma transfers are a common occurrence in rural areas, where critical access and lower-level trauma centers routinely transfer to tertiary care centers for specialized care. Transfers are non-therapeutic (NTT) when no specialist intervention occurs, leading to transfer that were futile (FT) or secondary overtriage (SOT). This study aimed to evaluate the prevalence of NTT among four trauma centers providing care to rural Appalachia. MethodsThis retrospective review was performed at four, ACS verified, Level 1 trauma centers. All adult trauma patients, transferred during 2018 were included for analysis. Transfers were considered futile if in <48 h the patient died or was discharged to hospice, without operative intervention. SOT transfers were discharged in <48 h, without major intervention, with an ISS< 15. Cost analysis was performed to describe the impact of NTT on EMS use. Results4,189 patients were analyzed during the study period. 105 (2.5%) met criteria for futility. Futile patients had a median ISS of 25 (IQR 9–26), and 48% had an AIS head ≥4. These were significantly greater (p<0.001) than non-futile transfers, median ISS 5 (IQR 2–9), 3% severe head injury. SOT occurred in 1371 (33%), median ISS of 5, and lower AIS scores by region. Isolated facial injuries resulted in 165 transfers. 13% of FT+SOT were admitted to the ICU. Only 22% of FT+SOT came from a trauma center. 68% were transported by ALS and 13% transported by air transport. FT+SOT traveled on average 70 miles from their home to receive care. ConclusionsNon-therapeutic transfers account for more than 1/3 of transfers in this rural environment. There was a significant use of advanced life support and aeromedical transport. The utility of these transfers should be questioned. With the recent increases in telehealth there is an opportunity for trauma systems to improve regional care and decrease transfers for futile cases.

Characterization of Hemagglutinin Gene Fragment of H9N2 Avian Influenza Virus Isolated from Environmental Live Bird Market in the Greater Jakarta Area

Virus H9N2 has become the most common subtype of low pathogenic avian influenza (LPAI) in poultry and shows the ability to infect humans. One of the important factors triggering the virus ​​pandemic is the live bird market (LBM). The virus acts as an internal gene donor in the other subtypes, such as H5N1, H5N2, H7N9, and H10N8, in poultry and humans. This study was conducted to detect the presence of the H9N2 virus and the molecular characteristics of the hemagglutinin gene fragment at the cleavage site, receptor binding site, antigenic site, and glycosylation site positions of the H9N2 subtype AI virus isolated from the LBM environment. This study used Disease Investigation Center Subang isolates from environmental samples of LBM in the Greater Jakarta area (DKI Jakarta, Bogor, Depok, Tangerang, and Bekasi) in 2019. Based on molecular detection using RT-PCR and RT-qPCR, it was found that avian influenza subtype H9N2 was detected. This indicates that LBM has the potential to be the source of the spread of the ​​H9N2 virus. The pattern of the amino acid cleavage site PSRSSR↓GLF indicates that the research sample belongs to the low pathogenic AI. The substitution of amino acids in the receptor binding site and antigenic site increases the specificity of non-human host recognition. The potential glycosylation site, with the NCS motif found in amino acid position 295-297, close to the receptor-binding site. Based on genetic analysis and phylogenetic topography, the virus is included in the CVI (China, Vietnam, Indonesia) lineage and subclade H9.4.2.5.

0639 Polysomnographic Biomarkers of Sleep Disruption and Sleep Disordered Breathing in Migraine: a Large Matched Case Control Clinical Registry-Based Study

Abstract Introduction Sleep disruption and sleep architectural changes in relation to migraine are not well characterized. Disruption in sleep may serve as a risk for migraine or result from migraine. We leveraged a large clinical cohort to examine the hypothesis that those with migraine have greater degrees of sleep architectural alterations and sleep disordered breathing (SDB). Methods This was a polysomnogram-based retrospective case (migraine) control (non-migraine) study of patients aged &amp;gt;18 matched 1:3 on age, sex, race, body mass index (BMI), and year of polysomnogram. Two domains were considered: 1)Sleep architecture (arousal index:AI, (primary predictor), total sleep time (TST), percentage of sleep stage time) and 2)SDB (apnea hypopnea index (AHI:primary predictor), mean oxygen saturation) were considered. Comparisons were performed by two-sample t-test or Wilcoxon rank sum test for continuous variables, and chi-square test or Fisher’s exact test for categorical variables. Results 4,783 migraine cases (47.5 ±13.3 years, 76.4% Caucasian, body mass index:BMI 33.7 ±8.6kg/m2) were matched to 14,287 controls. In migraine patients vs those without, TST was lower (359.0[307.0, 421.0] minutes vs 363.0[306.0, 432.5] minutes,p=0.01), percentage of N2 was higher (67.8%[59.6, 75.6] vs 67.0%[58.4, 74.8],p&amp;lt;0.001), percentage of REM was lower (16.7% [10.0, 22.0] vs 17.0% [11.1, 22.2],p=0.012), AHI was lower (7.4 [2.6, 17.0] vs 9.5 [3.7, 22.1],p&amp;lt;0.001), AI was lower (19.6 [12.8, 30.9] vs 22.6 [14.7, 34.9],p&amp;lt;0.001), and mean oxygen saturation was higher (93.7%±2.4 vs 93.3±2.6,p&amp;lt;0.001). Conclusion In this largest study of its kind, we identify novel associations of migraine in relation to curtailed sleep and sleep architectural alterations, i.e. increase in N2 and reduced REM sleep and lower AI compared to contemporaneously matched controls. Directionality of these relationships requires further elucidation given the cross-sectional nature of this study. Interestingly, we observed lower degree of sleep apnea and hypoxia burden in patients with migraine. As Calcitonin Gene-Related Peptide (CGRP), a neuropeptide increased during and between migraine attacks in migraine patients, and serotonin are implicated in arousals due to apnea-related increases in CO2, there is biologic plausibility for migraines patients to exhibit potential protection from SDB. Further investigation is needed to confirm these findings. Support (If Any)

AI-deploying organizations are key to addressing ‘perfect storm’ of AI risks

We argue that a perfect storm of five conditions heightens the risk of harm to society from artificial intelligence: (1) the powerful, invisible nature of AI, (2) low public awareness and AI literacy, (3) rapid scaled deployment of AI, (4) insufficient regulation, and (5) the gap between trustworthy AI principles and practices. To prevent harm, fit-for-purpose regulation and public AI literacy programs have been recommended, but education and government regulation will not be sufficient: AI-deploying organizations need to play a central role in creating and deploying trustworthy AI in line with the principles of trustworthy AI, and taking accountability to mitigate the risks.

Knowing When to Pass: The Effect of AI Reliability in Risky Decision Contexts.

This study manipulates the presence and reliability of AI recommendations for risky decisions to measure the effect on task performance, behavioral consequences of trust, and deviation from a probability matching collaborative decision-making model. Although AI decision support improves performance, people tend to underutilize AI recommendations, particularly when outcomes are uncertain. As AI reliability increases, task performance improves, largely due to higher rates of compliance (following action recommendations) and reliance (following no-action recommendations). In a between-subject design, participants were assigned to a high reliability AI, low reliability AI, or a control condition. Participants decided whether to bet that their team would win in a series of basketball games tying compensation to performance. We evaluated task performance (in accuracy and signal detection terms) and the behavioral consequences of trust (via compliance and reliance). AI recommendations improved task performance, had limited impact on risk-taking behavior, and were under-valued by participants. Accuracy, sensitivity (d'), and reliance increased in the high reliability AI condition, but there was no effect on response bias (c) or compliance. Participant behavior was only consistent with a probability matching model for compliance in the low reliability condition. In a pay-off structure that incentivized risk-taking, the primary value of the AI recommendations was in determining when to perform no action (i.e., pass on bets). In risky contexts, designers need to consider whether action or no-action recommendations will be more influential to design appropriate interventions.