5 considerations for building generative AI trust

Katie Schmidt, senior manager of AI/machine learning user experience research, contributed to this blog.

Years of user experience (UX) research and customer feedback have shown that building trust is integral to success with AI. With generative AI (GenAI), trust is more important than ever.

At ServiceNow, we know that when using GenAI capabilities, moments of potential trust touch all personas—from admins setting it up to developers using CodeAssist to agents using it to eliminate manual repetitive tasks and requesters seeking fast solutions to problems.

For example, agents use the Now Platform to help them solve tickets for requesters. There’s a tremendous opportunity to use GenAI capabilities to help agents work more productively and reduce mundane, repetitive tasks (that often lead to high turnover).

No one likes to feel like their purpose at work is to deal with frustrated, angry people or to copy and paste from a knowledge base article into a chat because a customer couldn’t find what they were looking for. GenAI has the potential to help agents focus on more meaningful work.

One way we’re making that reality is through expanded use cases for GenAI across all workflows. Using GenAI, our new case summarisation solution automatically recaps IT, HR and customer case information to help accelerate resolution times, minimise manual processes and boost agent productivity.
 

Keeping humans front and centre

Despite its many benefits, GenAI presents heightened concerns about security, regulations, data access and privacy, and cost, as well as overreliance on it in lieu of critical human oversight.

That’s why it’s imperative that we design and develop products using human-centred and responsible AI guidelines. Prioritising human experience will produce the desired GenAI impact for organisations’ employees and customers.

AI should augment and enhance how people work—but not fully automate their tasks or replace them. Based on UX research we’ve conducted with GenAI, we’ve come up with five key considerations to guide a human approach. These all tie into the broader goal of helping people build trust and feel in control.
 

1. Explainability

Assume people are uncertain about AI. Manage expectations, explain concepts and terms, and use experiences that are familiar to people where possible. People need to be able to understand and interpret predictions made by AI, particularly in the context of their work, where they’re accountable for the output. AI shouldn’t be “magic”.
 

2. Transparency

Overall, workers are very receptive to trying out AI at work but want to know when AI is the source. Clearly identify when AI is present. Users need to understand what AI did and why, and when necessary, users should be able to figure out how it works (e.g. data sources).
 

3. Ability to provide feedback

Humans recognise that AI isn’t perfect, but it’s important to be clear on its limits. This allows people to stay in control, provide feedback, and review the outputs from GenAI for accuracy. Let people mute, emphasise or redirect AI and its predictions. Make it easy to stop AI or to undo its actions quickly and with minimal to no harm.
 

4. Onboarding and building confidence

Your AI strategy should include onboarding, as change management is needed. Some onboarding and confidence building may be designed into the product, but the business offering the new UX must set expectations and take steps to help people adapt to it. Businesses should focus on quick wins to see early value.
 

5. Varied user acceptance

In our journey with AI at ServiceNow, we’ve seen the change in people's comfort, acceptance and reliance on it over time as they build trust and confidence in it. More people in various roles and industries can try to potentially benefit from GenAI, but they must see the value in it early on.

Recognise that people will approach the use of AI with different mindsets, characteristics, beliefs and levels of willingness to use it, particularly in work versus in their personal lives.

At ServiceNow, we always put people first. GenAI is no exception. We’re committed to learning and growing with our customers and providing human-centred AI best practices that prioritise building trust as we navigate this future.

As ServiceNow CEO Bill McDermott says, “Trust is the ultimate human currency. It’s earned in drops and lost in buckets.” We make sure our AI products continue to help build that trust, one drop at a time.

Find out more about ServiceNow generative AI. And sign up to participate in our AI user research.
 

© 2023 ServiceNow, Inc. All rights reserved. ServiceNow, the ServiceNow logo, Now and other ServiceNow marks are trademarks and/or registered trademarks of ServiceNow, Inc. in the United States and/or other countries. Other company names, product names and logos may be trademarks of the respective companies with which they are associated.