The Secret Ingredient for Building Trust in AI
When creating AI products, a fundamental question arises: how do we get users to truly trust these advanced technologies?
Recent research reveals a crucial insight into building trust in AI systems: while performance metrics matter, emotional connections serve as the cornerstone for sustainable user engagement.
This study examining artificial intelligence implementation in the workplace demonstrates how stakeholders develop both cognitive and emotional bonds with AI tools.
Cognitive Trust as the foundational layer
The research highlights how users form cognitive trust through clear outputs and responsible AI implementation. This trust develops through three key elements:
System reliability in delivering consistent, explainable outputs
Transparency in AI algorithms and decision-making processes
Demonstrated competence in providing valuable stakeholder solutions
However, when users only develop cognitive trust, they often limit their AI adoption to narrow, specific use cases. For instance, employees might restrict their interaction to basic information retrieval tasks, showing reluctance to explore broader AI applications.
For example
Users who only had cognitive trust used the chatbot for the same type of simple information retrieval task, like looking up HR records.
They did not try expanding the use of the chatbot to new contexts or questions.
Their cognitive trust was contingent on it performing that one function well.
The Critical Role of Emotional Trust
The study reveals that implementing safeguards and ensuring explainability alone aren't sufficient for building trust in artificial intelligence. Emotional connections emerge through:
In contrast, emotional trust emerged from something else:
1. Anthromorphic design
Anthropomorphic design touches like names, avatars, and messaging that created rapport and affinity.
“Employees experienced emotions towards the AI system, stemming from their feelings of closeness to the chatbot. Indeed, employees expressed affection towards its icon and animation”
“The tangibility of the embodiment of the AI chatbot, i.e. the visual representation of the AI chatbot, its name and its identity as a gender-neutral ‘funny creature’, allowed users to connect with the AI chatbot and build a personal bond with it.”
2. Sense of care
A sense of care and partnership between the user and the machine.
“They felt that interacting with the AI chatbot gave them the feeling that someone was listening to them and that this resembled human–human communication.”
3. Human-like connection
Feeling an intuitive, human-like connection with the technology:
“The personal messages that accompany the AI chatbot consistently remind the user that ‘Brainy is learning’ and that employees could ‘teach the bot’ if they choose to."
This emotional glue caused users to persistently experiment with the chatbot across new contexts. Despite imperfections, emotional investment paid dividends in forgiveness and patience, generating data to enhance its learning.
The Synergy: Cognitive and Emotional Trust in AI Governance
The study revealed cognitive and emotional factors mutually reinforce one another to build holistic user trust.
Cognitive trust in capabilities provides the bedrock baseline functionality people expect from AI. But emotional trust sustains engagement beyond early limitations, enabling rapid iteration.
Without this emotional foundation, people abandon tools at the first fault. But with care invested, they grow more tolerant of missteps, granting time for progress.
While AI algorithms and data quality provide the foundation, emotional investment sustains engagement through inevitable challenges in the AI lifecycle.
This insight carries significant implications for AI development and implementation:
Focus on both explainability and emotional connection
Balance clear outputs with engaging interactions
Implement safeguards while maintaining approachability
Cognitive Trust + Emotional Trust = Comprehensive Trust
Constructing comprehensive trust requires harmonizing logical and emotional elements.
To create lasting trusted AI systems, organizations should:
Establish a functional baseline of reliability, transparency, and competence that users can logically trust. (Supported by findings on cognitive trust factors).
Infuse emotional design - personalities, dialogue, rapport - to build intuitive bonds. (As evidenced by emotional trust driven by anthropomorphism and reciprocity).
Allow space for imperfections while enhancing capabilities over time. (Emotional trust sustained engagement despite limitations, enabling improvement).
Cultivate a partnership mindset between the user and the AI while maintaining high standards for data protection and privacy. (Findings showed a sense of cooperation and teamwork).
Bridging the logical and emotional, we can design an AI solution that aligns with both the head and the heart, cultivating resilient user trust.
Thanks for reading!
Link to the study:
2023, Gkinko, L., Elbanna, A. “Designing trust: The formation of employees’ trust in conversational AI in the digital workplace” - Link to ScieneDirect