How Human Intelligence is Affected by Artificial Intelligence: Recent Studies
Recent studies have shed light on the relationship between human intelligence and artificial intelligence (AI). These studies offer valuable insights into the biases and perceptions that can arise when interacting with AI. This article will discuss two recent research studies that explore these topics, one conducted by Deusto University and another by MIT and Arizona State University.
The Deusto University Study: Do We Catch Biases from AI Bots?
A study conducted by Helena Matute and Lucia Vicente from Deusto University highlighted the potential biases and errors associated with AI. The researchers emphasized that AI systems like ChatGPT provide results based on their training, which means that any biases or errors present in their training data can manifest in the outputs they produce. This has significant implications as we rely on these AI systems for various applications, which can result in unintended consequences if we are not careful.
Matute and Vicente conducted three experiments to test this hypothesis. Participants were asked to perform a medical diagnosis task, with one group receiving assistance from a biased AI system that exhibited systematic errors, while the other group served as the control group without AI assistance. The researchers used a fictitious AI system, medical diagnosis task, and disease to ensure that the experiments did not interfere with real situations.
The results of the study showed that the group assisted by the biased AI system made the same errors as the AI system itself, while the control group did not. This suggests that AI recommendations can influence human decisions. Even when the AI assistance was removed, the group that received assistance continued to make the same systematic errors. In contrast, the control group remained unaffected by the AI group’s decisions. These findings highlight the potential negative impact of AI biases on human decision-making, although further research is needed to confirm the inheritance of AI biases by individuals.
The MIT and Arizona State University Study: How Biases Alter our Perception of AI
In contrast to the Deusto University study, a study conducted by researchers from MIT and Arizona State University focused on how our preconceived notions about AI can affect our perception of AI outputs. The researchers aimed to understand the effect of AI placebo and how our imagination of AI shapes our interaction with it.
To test this hypothesis, 300 volunteers were divided into three groups. Each group was assigned a specific role: the first group was told that the AI had no ulterior motives and was simply a mundane text completion program, the second group was told the AI had empathy, and the third group was told the AI was manipulative and only interested in selling a service. However, all groups interacted with the same AI program.
After speaking with the AI for 10 to 30 minutes, the participants were asked to evaluate whether the AI was an effective mental health companion. The results showed that only 44% of those who received negative primers believed their assigned role, while 79% of the neutral group considered the AI to be neutral, and 88% of the positive group perceived the AI as empathetic. These findings emphasize the influence of our preconceived notions on how we perceive and interact with AI systems.
The senior author of the study, Pattie Maes, suggested that we should educate people to be more cautious and aware of the potential biases and hallucinations of AI systems. How we communicate and discuss AI systems will ultimately shape people’s responses to them.
Conclusion
These recent studies provide important insights into the impact of AI on human intelligence. The study from Deusto University emphasizes the potential biases that can be passed on to individuals through AI, while the study from MIT and Arizona State University highlights how our preconceived notions about AI can shape our perception of AI outputs. It is crucial for us to understand and address these biases and perceptions to ensure responsible and effective use of AI systems in our daily lives.
Denial of responsibility! Vigour Times is an automatic aggregator of Global media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, and all materials to their authors. For any complaint, please reach us at – [email protected]. We will take necessary action within 24 hours.
