Artificial intelligence chatbots exhibits similar biases to humans, according to new research published in Proceedings of the National Academy of Sciences of the United States of America (PNAS). The study suggests that AI tends to favor certain types of information over others, reflecting patterns seen in human communication.
The motivation behind this research lies in the burgeoning influence of large language models like ChatGPT-3 in various fields. With the wide application of these AI systems, understanding how they might replicate human biases becomes crucial.
Observatorio IA - sesgos
New research by the psychologists Lucía Vicente and Helena Matute from Deusto University in Bilbao, Spain, provides evidence that people can inherit artificial intelligence biases (systematic errors in AI outputs) in their decisions.
The astonishing results achieved by artificial intelligence systems, which can, for example, hold a conversation as a human does, have given this technology an image of high reliability. More and more professional fields are implementing AI-based tools to support the decision-making of specialists to minimise errors in their decisions. However, this technology is not without risks due to biases in AI results. We must consider that the data used to train AI models reflects past human decisions. If this data hides patterns of systematic errors, the AI algorithm will learn and reproduce these errors. Indeed, extensive evidence indicates that AI systems do inherit and amplify human biases.
The most relevant finding of Vicente and Matute’s research is that the opposite effect may also occur: that humans inherit AI biases. That is, not only would AI inherit its biases from human data, but people could also inherit those biases from AI, with the risk of getting trapped in a dangerous loop. Scientific Reports publishes the results of Vicente and Matute’s research.
In the series of three experiments conducted by these researchers, volunteers performed a medical diagnosis task. A group of the participants were assisted by a biased AI system (it exhibited a systematic error) during this task, while the control group were unassisted. The AI, the medical diagnosis task, and the disease were fictitious. The whole setting was a simulation to avoid interference with real situations.
The participants assisted by the biased AI system made the same type of errors as the AI, while the control group did not make these mistakes. Thus, AI recommendations influenced participant’s decisions. Yet the most significant finding of the research was that, after interaction with the AI system, those volunteers continued to mimic its systematic error when they switched to performing the diagnosis task unaided. In other words, participants who were first assisted by the biased AI replicated its bias in a context without this support, thus showing an inherited bias. This effect was not observed for the participants in the control group, who performed the task unaided from the beginning.
These results show that biased information by an artificial intelligence model can have a perdurable negative impact on human decisions. The finding of an inheritance of AI bias effect points to the need for further psychological and multidisciplinary research on AI-human interaction. Furthermore, evidence-based regulation is also needed to guarantee fair and ethical AI, considering not only the AI technical features but also the psychological aspects of the IA and human collaboration.
Que los humanos introducimos nuestros sesgos en los algoritmos no es ninguna novedad y es una de las principales fuentes de preocupación sobre estas tecnologías. Lo que no sabíamos hasta ahora es que el proceso puede suceder a la inversa: la máquina también puede contagiarnos a nosotros sus errores sistemáticos y estos se pueden perpetuar. Es lo que concluyen las investigadoras españolas Lucía Vicente y Helena Matute después de realizar una serie de experimentos cuyos resultados publican este martes en la revista Scientific Reports.
Tipo
-
aplicación (4)
-
artículo (86)
-
artículo científico (12)
-
boletín (3)
-
curso (1)
-
libro (2)
-
podcast (1)
-
presentación (6)
-
revista (1)
-
sitio web (11)
-
tuit (26)
-
vídeo (7)
Temas
-
aplicaciones (30)
-
aprendizaje (8)
-
Australia (1)
-
Bard (4)
-
bibliografía (1)
-
big data (1)
-
Bing (3)
-
brecha digital (1)
-
chatbots (8)
-
chatGPT (44)
-
código abierto (1)
-
cognición (1)
-
cómo citar (2)
-
comparación (3)
-
consejos de uso (2)
-
control (1)
-
curso (1)
-
cursos (1)
-
DALLE2 (1)
-
deep learning (2)
-
deepfakes (1)
-
desafíos (1)
-
destrezas (1)
-
detección de uso (6)
-
diseño educativo (1)
-
disrupción (1)
-
docentes (1)
-
e-learning (1)
-
economía (1)
-
educación (48)
-
educación superior (10)
-
educadores (1)
-
embeddings (1)
-
encuesta (2)
-
enseñanza (3)
-
enseñanza de IA (1)
-
entrevista (2)
-
escritura (2)
-
evaluación (4)
-
experimentación (1)
-
futuro de la IA (2)
-
futuro laboral (1)
-
Google (2)
-
Google Docs (1)
-
guía (1)
-
guía de uso (5)
-
historia (1)
-
IA generativa (3)
-
IA vs. humanos (1)
-
ideas (1)
-
ideas de uso (7)
-
impacto social (1)
-
información (1)
-
investigación (5)
-
Kahoot (1)
-
LLM (2)
-
machine learning (2)
-
mapas (1)
-
medicina (1)
-
Microsoft (1)
-
Midjourney (1)
-
mundo laboral (1)
-
música (1)
-
niños (1)
-
noticias (1)
-
OpenAI (1)
-
opinión (4)
-
orígenes (1)
-
pedagogía (1)
-
plagio (3)
-
plugins (2)
-
presentación (1)
-
problemas (2)
-
programación (1)
-
prompts (1)
-
recomendaciones (1)
-
recopilación (15)
-
recursos (1)
-
regulación (3)
-
revista (1)
-
riesgos (5)
-
robots (1)
-
sesgos (3)
-
trabajo (3)
-
traducción (2)
-
turismo (1)
-
tutorbots (1)
-
tutores de IA (2)
-
tutoriales (3)
-
uso de la lengua (1)
-
uso del español (1)
-
uso en educación (6)
-
usos (4)
-
valoración (1)
-
viajes (1)
Autores
-
A. Lockett (1)
-
AI Foreground (1)
-
Alejandro Tinoco (1)
-
Alfaiz Ali (1)
-
Anca Dragan (1)
-
Andrew Yao (1)
-
Anna Mills (1)
-
Antonio Byrd (1)
-
Ashwin Acharya (1)
-
Barnard College (1)
-
Barsee (1)
-
Ben Dickson (1)
-
Brian Basgen (1)
-
Brian Roemmele (1)
-
Brian X. Chen (4)
-
Carmen Rodríguez (1)
-
Carrie Spector (1)
-
Ceren Ocak (1)
-
Ceylan Yeginsu (1)
-
Charles Hodges (1)
-
Csaba Kissi (1)
-
Daniel Kahneman (1)
-
David Álvarez (1)
-
David Green (1)
-
David Krueger (1)
-
Dawn Song (1)
-
DeepLearning.AI (1)
-
Dennis Pierce (1)
-
Dimitri Kanaris (1)
-
Eli Collins (1)
-
Emily Bender (1)
-
Enrique Dans (3)
-
Eric M. Anderman (1)
-
Eric W. Dolan (1)
-
Eric Wu (1)
-
Ethan Mollick (1)
-
Eva M. González (1)
-
Francis Y (3)
-
Frank Hutter (1)
-
Gary Marcus (1)
-
Geoffrey Hinton (1)
-
George Siemens (3)
-
Gillian Hadfield (1)
-
Gonzalo Abio (1)
-
Google (3)
-
Gorka Garate (1)
-
Greg Brockman (1)
-
Guillaume Bardet (1)
-
Hasan Toor (4)
-
Hassan Khosravi (1)
-
Helen Beetham (1)
-
Helena Matute (1)
-
Hélène Sauzéon (1)
-
Holly Hassel (1)
-
Ian Roberts (1)
-
James Zou (1)
-
Jan Brauner (1)
-
Jas Singh (3)
-
Javier Pastor (1)
-
Jeff Clune (1)
-
Jeffrey Watumull (1)
-
Jenay Robert (1)
-
Johanna C. (1)
-
Johannes Wachs (1)
-
Josh Bersin (1)
-
Juan Cuccarese (1)
-
Julian Estevez (1)
-
Kalley Huang (1)
-
Karie Willyerd (1)
-
Kevin Roose (1)
-
Kui Xie (1)
-
Lan Xue (1)
-
Lance Eaton (1)
-
Leonardo Flores (1)
-
Lijia Chen (1)
-
Lorna Waddington (1)
-
Lucía Vicente (1)
-
Manuel Graña (1)
-
Mark McCormack (1)
-
Marko Kolanovic (1)
-
Melissa Heikkilä (1)
-
Mert Yuksekgonul (1)
-
Microsoft (1)
-
MLA Style Center (1)
-
Muzzammil (1)
-
Nada Lavrač (1)
-
Naomi S. Baron (1)
-
Natasha Singer (2)
-
Nathan Lands (1)
-
Nicole Muscanell (1)
-
Nikki Siapno (1)
-
NLLB Team (1)
-
Noam Chomsky (1)
-
Nuria Oliver (1)
-
Oliver Whang (1)
-
Olumide Popoola (1)
-
OpenAI (2)
-
Paul Couvert (5)
-
Paula Escobar (1)
-
Pauline Lucas (1)
-
Petr Šigut (1)
-
Philip Torr (1)
-
Philippa Hardman (18)
-
Pieter Abbeel (1)
-
Pingping Chen (1)
-
Pratham (1)
-
Qiqi Gao (1)
-
Rafael Ruiz (1)
-
Rania Abdelghani (1)
-
Rebecca Marrone (1)
-
Rishit Patel (1)
-
Rowan Cheung (2)
-
Russell Group (1)
-
Sal Khan (1)
-
Samuel A. Pilar (1)
-
Samuel Fowler (1)
-
Sarah Z. Johnson (1)
-
Sepp Hochreiter (1)
-
Serge Belongie (1)
-
Shazia Sadiq (1)
-
Sheila McIlraith (1)
-
Sihem Amer-Yahia (1)
-
Sonja Bjelobaba (1)
-
Sören Mindermann (1)
-
Stan Waddell (1)
-
Stella Tan (1)
-
Stephen Marche (1)
-
Steve Lohr (1)
-
Stuart Russell (1)
-
Tegan Maharaj (1)
-
Tiffany Hsu (1)
-
Tim Leberecht (1)
-
Timothy McAdoo (1)
-
Tom Graham (1)
-
Tom Warren (1)
-
Tomáš Foltýnek (1)
-
Tong Wang (1)
-
Trevor Darrell (1)
-
Tulsi Soni (2)
-
Vicki Boykis. (1)
-
Víctor Millán (1)
-
Weixin Liang (1)
-
Xingdi Yuan (1)
-
Ya-Qin Zhang (1)
-
Yejin Choi (1)
-
Yen-Hsiang Wang (1)
-
Yining Mao (1)
-
Yoshua Bengio (1)
-
Yurii Nykon (1)
-
Zhijian Lin (1)
Fuentes
-
APA Style (1)
-
arXiv (4)
-
E-aprendizaje.es (1)
-
EDUCAUSE (7)
-
Educaweb (1)
-
El País (1)
-
ElDiario.es (3)
-
Enrique Dans (1)
-
eSchool News (1)
-
Formación ELE (1)
-
Generación EZ (1)
-
GP Strategies (1)
-
HigherEdJobs (1)
-
IE Insights (1)
-
IEEE Access (2)
-
INTEF (1)
-
Intellias (1)
-
J.P.Morgan (1)
-
Joshbersin.com (1)
-
Kahoot! (1)
-
La Tercera (1)
-
Learning Letters (2)
-
Medium (1)
-
Meta AI (1)
-
Meta Research (1)
-
MLA (1)
-
Multiplex (1)
-
New York Times (14)
-
Open AI (1)
-
OpenAI (2)
-
PsyPost (1)
-
RTVE (1)
-
Russell Group (1)
-
Science (1)
-
TED (5)
-
TEDx (1)
-
The Atlantic (1)
-
The Conversation (4)
-
The Rundown (1)
-
The Verge (1)
-
ThinkBig (1)
-
Twitter (26)
-
Xataca (1)
-
Youtube (6)