Artificial intelligence and illusions of understanding in scientific research

Nature, Journal Year: 2024, Volume and Issue: 627(8002), P. 49 - 58

Published: March 6, 2024

Language: Английский

---

Artificial intelligence in healthcare: Complementing, not replacing, doctors and healthcare providers

Digital Health, Journal Year: 2023, Volume and Issue: 9

Published: Jan. 1, 2023

The utilization of artificial intelligence (AI) in clinical practice has increased and is evidently contributing to improved diagnostic accuracy, optimized treatment planning, patient outcomes. rapid evolution AI, especially generative AI large language models (LLMs), have reignited the discussions about their potential impact on healthcare industry, particularly regarding role providers. Concerning questions, “can replace doctors?” “will doctors who are using those not it?” been echoed. To shed light this debate, article focuses emphasizing augmentative healthcare, underlining that aimed complement, rather than replace, fundamental solution emerges with human–AI collaboration, which combines cognitive strengths providers analytical capabilities AI. A human-in-the-loop (HITL) approach ensures systems guided, communicated, supervised by human expertise, thereby maintaining safety quality services. Finally, adoption can be forged further organizational process informed HITL improve multidisciplinary teams loop. create a paradigm shift complementing enhancing skills providers, ultimately leading service quality, outcomes, more efficient system.

Language: Английский

---

Dissociating language and thought in large language models

Trends in Cognitive Sciences, Journal Year: 2024, Volume and Issue: 28(6), P. 517 - 540

Published: March 19, 2024

Language: Английский

---

The Janus Effect of Generative AI: Charting the Path for Responsible Conduct of Scholarly Activities in Information Systems

Information Systems Research, Journal Year: 2023, Volume and Issue: 34(2), P. 399 - 408

Published: May 26, 2023

Funding: A. Susarla was funded by an R01 grant from the National Library of Medicine, through [Grant R01LM013443].

Language: Английский

---

The feasibility of artificial consciousness through the lens of neuroscience

Trends in Neurosciences, Journal Year: 2023, Volume and Issue: 46(12), P. 1008 - 1017

Published: Oct. 18, 2023

Language: Английский

---

The emergence of economic rationality of GPT

Proceedings of the National Academy of Sciences, Journal Year: 2023, Volume and Issue: 120(51)

Published: Dec. 12, 2023

As large language models (LLMs) like GPT become increasingly prevalent, it is essential that we assess their capabilities beyond processing. This paper examines the economic rationality of by instructing to make budgetary decisions in four domains: risk, time, social, and food preferences. We measure assessing consistency GPT’s with utility maximization classic revealed preference theory. find are largely rational each domain demonstrate higher score than those human subjects a parallel experiment literature. Moreover, estimated parameters slightly different from exhibit lower degree heterogeneity. also scores robust randomness demographic settings such as age gender but sensitive contexts based on frames choice situations. These results suggest potential LLMs good need further understand capabilities, limitations, underlying mechanisms.

Language: Английский

---

Personality Traits in Large Language Models

Research Square (Research Square), Journal Year: 2023, Volume and Issue: unknown

Published: Aug. 28, 2023

Abstract The advent of large language models (LLMs) has revolutionized natural processing, enabling the generation coherent and contextually relevant text. As LLMs increasingly power conversational agents, synthetic personality embedded in these models, by virtue training on amounts human data, is becoming important. Since a key factor determining effectiveness communication, we present comprehensive method for administering validating tests widely-used LLMs, as well shaping generated text such LLMs. Applying this method, found: 1) measurements outputs some under specific prompting configurations are reliable valid; 2) evidence reliability validity LLM stronger larger instruction fine-tuned models; 3) can be shaped along desired dimensions to mimic profiles. We discuss application ethical implications measurement particular regarding responsible use

Language: Английский

---

Can Generative AI improve social science?

Proceedings of the National Academy of Sciences, Journal Year: 2024, Volume and Issue: 121(21)

Published: May 9, 2024

Generative AI that can produce realistic text, images, and other human-like outputs is currently transforming many different industries. Yet it not yet known how such tools might influence social science research. I argue has the potential to improve survey research, online experiments, automated content analyses, agent-based models, techniques commonly used study human behavior. In second section of this article, discuss limitations AI. examine bias in data train these negatively impact research—as well as a range challenges related ethics, replication, environmental impact, proliferation low-quality conclude by arguing scientists address creating open-source infrastructure for research on Such only necessary ensure broad access high-quality tools, argue, but also because progress will require deeper understanding forces guide

Language: Английский

---

CancerGPT for few shot drug pair synergy prediction using large pretrained language models

npj Digital Medicine, Journal Year: 2024, Volume and Issue: 7(1)

Published: Feb. 19, 2024

Large pre-trained language models (LLMs) have been shown to significant potential in few-shot learning across various fields, even with minimal training data. However, their ability generalize unseen tasks more complex such as biology, has yet be fully evaluated. LLMs can offer a promising alternative approach for biological inference, particularly cases where structured data and sample size are limited, by extracting prior knowledge from text corpora. Our proposed uses predict the synergy of drug pairs rare tissues that lack features. experiments, which involved seven different cancer types, demonstrated LLM-based prediction model achieved accuracy very few or zero samples. model, CancerGPT (with ~ 124M parameters), was comparable larger fine-tuned GPT-3 175B parameters). research is first tackle pair limited We also utilize an reaction tasks.

Language: Английский

---

Cognitive Network Science Reveals Bias in GPT-3, GPT-3.5 Turbo, and GPT-4 Mirroring Math Anxiety in High-School Students

Big Data and Cognitive Computing, Journal Year: 2023, Volume and Issue: 7(3), P. 124 - 124

Published: June 27, 2023

Large Language Models (LLMs) are becoming increasingly integrated into our lives. Hence, it is important to understand the biases present in their outputs order avoid perpetuating harmful stereotypes, which originate own flawed ways of thinking. This challenge requires developing new benchmarks and methods for quantifying affective semantic bias, keeping mind that LLMs act as psycho-social mirrors reflect views tendencies prevalent society. One such tendency has negative effects global phenomenon anxiety toward math STEM subjects. In this study, we introduce a novel application network science cognitive psychology towards fields from ChatGPT, GPT-3, GPT-3.5, GPT-4. Specifically, use behavioral forma mentis networks (BFMNs) how these frame disciplines relation other concepts. We data obtained by probing three language generation task previously been applied humans. Our findings indicate have perceptions fields, associating with concepts 6 cases out 10. observe significant differences across OpenAI’s models: newer versions (i.e., GPT-4) produce 5× semantically richer, more emotionally polarized fewer associations compared older N=159 high-school students. These suggest advances architecture may lead less biased models could even perhaps someday aid reducing stereotypes society rather than them.

Language: Английский

---