How can businesses build greater trust in AI? With a human at the helm

Authored by Rowena Westphalen, Senior vice president innovation, AI & Customer Advisory APAC at Salesforce

Artificial intelligence (AI) has become one of the top priorities among many CEOs and senior leaders. It can help businesses build stronger customer relationships and support employees to work more efficiently, which both influence revenue and business resilience. To use AI to its full potential, businesses need to get buy-in from their staff and customers. So, how do you build trust in AI among your stakeholders? 

In Australia, our research shows that only 59% of Australian office workers trust generative AI. Another study found that over half of office workers are concerned about the data that’s used to train AI systems, with almost three-quarters (72%) believing that the training data underpinning AI is unreliable. 

As with any emerging technology, some apprehension is to be expected. We’re still learning what AI is fully capable of doing for businesses and society at large. When it comes to solving these challenges and demonstrating to stakeholders that your business is in control of AI, there is one key component to consider – making sure a human remains at the helm.

AI is continuing to evolve, so it’s essential businesses find the right balance between supercharging productivity and prioritising staff expertise. Almost nine in ten office workers don’t trust AI that operates without human guidance, and a further 90% feel more assured about AI and data security if a human is involved. It’s clear that the combination of human creativity and judgement with machine speed and efficiency will unlock productivity, empower employees, and ultimately, inspire trust in AI. 

So, how do businesses put a framework in place to build trust in AI both inside its walls and with customers? 

1. Define and agree on ethical principles and guidelines that fit your specific business goals and needs
2. Set up a diverse team to regularly review risks. For example, at Salesforce, we have the Office of Ethical and Humane Use of Technology to address questions and concerns surrounding the use of technologies like AI
3. Educate and boost employee confidence through workplace training. When your staff are comfortable using generative AI, get them involved with spotting bias and reducing risks – being proactive and keeping data safe should be a company-wide effort
4. Maintain confidentiality and security by developing large language models (LLMs) that are trained to write like humans and don’t retain data

Building trust in your AI data is essential for customer confidence and successful implementation. Customers want to know their data is secure and used ethically in today’s AI landscape of data theft and incorrect, or fabricated information.

This trust extends to the technology being used too, and highlights the importance of prioritising transparency and responsible data practices. 

• Define clear metrics that align with your customer value principles and identify use cases for AI across departments. 
• Ensure a unified data environment by connecting customer data and profiles to create a single source of truth. 
• Prioritise clean data, free from errors and inconsistencies. 
• Finally, implement robust security measures like data masking and zero-retention policies to safeguard data.

By taking these steps, you can build trust, meet customer expectations, and unlock AI’s full potential in your business.