Why AI-Based Guidance Could Foster Biases

Research has found that guidance generated by AI can shape human judgment in ways that may reinforce bias and weaken decision-making, particularly among people who already hold favorable views of AI systems. 

The study, published in the journal Scientific Reports, examined how people respond to guidance when making judgments, and whether the source of that guidance matters. According to the authors, a widespread belief that technology is objective and free from human bias may encourage users to rely too heavily on AI systems, even when those systems are flawed. 

To explore how people respond to AI advice, the researchers recruited 295 participants for an experiment focused on facial recognition. Participants were shown 80 images of faces and asked to decide whether each image depicted a real person or a face generated by AI. 

Alongside each image, participants were shown written guidance that claimed to offer expert predictions about the image. Some participants were told the guidance came from human specialists in facial recognition. Others were told it was produced by an AI system trained to distinguish real faces from artificial ones. 

Crucially, the guidance was intentionally unreliable. Unknown to the participants, it was accurate only 50% of the time. For both real and synthetic faces, half were paired with correct advice and half with incorrect advice. The order of images was randomized, and participants were not told that the accuracy of the guidance varied. 

After completing the task, participants filled out two questionnaires. One measured their general tendency to trust other people, while the other assessed their overall attitudes toward artificial intelligence. 

Participants who held more positive views of AI were less able to tell real faces from synthetic ones, but only when they received guidance attributed to AI. This decline in accuracy did not occur when the same participants were given advice said to come from human sources. 

The authors say that the findings raise concerns as AI tools are increasingly marketed as aids for complex decisions in real-world settings. They note that AI-based support may have a unique capacity to introduce bias, sometimes reducing rather than improving the quality of human judgment. People who are enthusiastic about AI, they add, may be especially at risk of placing too much trust in its outputs. 

The study concludes that further work is needed to better understand how people interpret and apply AI guidance across different situations, particularly as such systems become more embedded in daily life. 

As even more advanced technologies are commercialized by entities like D-Wave Quantum Inc. (NYSE: QBTS), the need to understand the extent to which technology can be relied upon and where human decision-making is required is critical. 

About AINewsWire

AINewsWire (“AINW”) is a specialized communications platform with a focus on the latest advancements in artificial intelligence (“AI”), including the technologies, trends and trailblazers driving innovation forward. It is one of 75+ brands within the Dynamic Brand Portfolio @ IBN that delivers: (1) access to a vast network of wire solutions via InvestorWire to efficiently and effectively reach a myriad of target markets, demographics and diverse industries; (2) article and editorial syndication to 5,000+ outlets; (3) enhanced press release enhancement to ensure maximum impact; (4) social media distribution via IBN to millions of social media followers; and (5) a full array of tailored corporate communications solutions. With broad reach and a seasoned team of contributing journalists and writers, AINW is uniquely positioned to best serve private and public companies that want to reach a wide audience of investors, influencers, consumers, journalists, and the general public. By cutting through the overload of information in today’s market, AINW brings its clients unparalleled recognition and brand awareness.

AINW is where breaking news, insightful content and actionable information converge.

To receive SMS alerts from AINewsWire, text “AI” to 888-902-4192 (U.S. Mobile Phones Only)

For more information, please visit www.AINewsWire.com

Please see full terms of use and disclaimers on the AINewsWire website applicable to all content provided by AINW, wherever published or re-published: https://www.AINewsWire.com/Disclaimer

AINewsWire
Los Angeles, CA
www.AINewsWire.com
310.299.1717 Office
Editor@AINewsWire.com

AINewsWire is powered by IBN