People perform better when they think AI is helping them

In the ever-evolving world of artificial intelligence (AI), we often find ourselves marveling at the latest advancements, from self-driving cars to language models that seemingly hold the world’s knowledge. But what if the true power of AI lies not just in its algorithms and data but also in our very belief in it? A fascinating new study from Aalto University in Finland suggests that just believing an AI is helping us can significantly boost our performance, even when the AI is unreliable or nonexistent.

AI impacts how we perform

Researchers at Aalto University created an innovative experiment to explore how our beliefs about AI affect our performance. Participants underwent a basic letter recognition exercise twice. They first completed it alone and then with supposed help from an AI system.

Adding a twist, the researchers divided the participants into two groups. One group was informed that the AI was reliable and would improve their performance. The other group was told the AI was unreliable and would degrade their performance.

The catch? Neither AI system actually existed. The researchers created a “sham-AI” that provided random feedback, regardless of the participants’ actual performance.

Belief trumps reality

The results of the study were unexpected and remarkable. Participants in both groups, regardless of whether they were informed of the AI’s unreliability or not, demonstrated a significant improvement in their performance on the letter recognition task when they believed they were being assisted by an AI system.

This improvement was evident in both the speed and accuracy with which they completed the task, indicating a heightened level of focus and engagement when participants perceived AI involvement.

“What we discovered is that people have extremely high expectations of these systems, and we can’t make them AI doomers simply by telling them a program doesn’t work,” explained Professor Robin Welsch. The researchers replicated the study online with similar results, further solidifying the power of belief in AI.

AI placebo effect

This phenomenon, dubbed the “AI placebo effect,” has profound implications for how we evaluate and interact with AI systems. “This is the big realization coming from our study – that it’s hard to evaluate programs that promise to help you because of this placebo effect,” said Welsch.

The AI placebo effect presents several challenges. Evaluating AI systems can be challenging due to the placebo effect. Firstly, this phenomenon makes it hard to discern if improvements in performance stem from the actual capabilities of the AI or merely from people’s belief in its effectiveness. Thus, distinguishing the true impact of AI technology requires careful consideration of how beliefs influence outcomes.

In the human-computer interaction research, expectations play a critical role in shaping study results. The anticipatory beliefs of participants can significantly sway the outcomes, potentially skewing the data. To address this, implementing placebo-controlled studies is crucial to isolate the real effects of AI from the psychological impact of expectations.

Marketing strategies often capitalize on the placebo effect, promoting AI systems in a way that heightens expectations among users. This approach can lead to exaggerated beliefs about the effectiveness of AI, which may not accurately reflect its actual capabilities. As a result, consumers may perceive AI technologies as more powerful than they are, complicating the assessment of their true utility and performance.

Belief in AI for better performance

While the AI placebo effect raises concerns, it also highlights the incredible power of our beliefs and expectations in shaping our experiences and how we perform. As AI continues to permeate our lives, understanding this phenomenon will be crucial for both researchers and users alike.

“People perform better if they think they have an AI assistant – even when they’ve been told it’s unreliable and won’t help them,” said doctoral researcher Agnes Kloft.

As we navigate the ever-evolving landscape of AI, it’s clear that our belief in technology plays a significant role in shaping its impact on our lives. The AI placebo effect is a fascinating reminder of the power of the human mind and the complex interplay between technology and our perceptions.

The research was presented at The ACM Conference on Human Factors in Computing Systems.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates. 

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–