“Quality built the industrial world. Trust will build the intelligent one.”
Guess who said the above quote. Yep. You guessed it! ChatGPT came up with the nifty quote. And it’s a great start for this piece.
The challenge we quality professionals are facing is AI is being integrated into all products and services. And this is a problem. We don’t know how the black box makes decisions, how our data is being used, and what human recourse we have to correct unfair autonomous decisions.
Quality and AI Coming Together
Trust and safety have always been part of quality, specifically quality assurance from ISO 9001 to today’s risk management. Quality trust and AI trust are now coming together – quickly. Product quality and service quality are becoming AI quality. So, let’s start at the beginning.
Most physical products have an internet connection and an AI component. Physical products are connected to the internet to monitor, check, and even surveil. Heart monitors are connected to doctor’s offices for 100% monitoring. Cars—computers on four wheels—have an internet connection so software is updated regularly. Traffic cameras surveil speeding motorists. The list goes on.
The same goes for service quality. Most service interactions incorporate AI in terms of customer service, chat, or search. We’ve all experienced these. For humans to use these AI systems, trust has to be baked into the service and product offerings.
In this piece, we look at how AI product and service quality are coming together to form new definitions of quality and its taxonomy—with huge implications for our profession.
Why Trust Matters
Quality at its most basic is all about trust. Trust communicates brand value. Trust assures that products conform to requirements. Trust implies that products are safe. Trust means that the human and bot customer service interaction solves our problem and satisfies us. And verifying and validating trust is the essence of quality assurance.
AI also assumes trust. Trust implies AI system decisions are explainable. AI trust means the system is safe, fair and not biased.
There are a lot of parallels to AI assurance and quality assurance. Let’s first look at dimensions that define quality and then the attributes that define AI.
Quality Dimensions
David Garvin, Harvard Business School professor, wrote in 1987 the seminal quality article of the last 50 years. Garvin identified 8 dimensions of quality. They are: performance, features, reliability, conformance, durability, serviceability, aesthetics, and perceived quality. These quality dimensions are audited to ensure trust.
Garvin’s quality dimensions became the vocabulary and taxonomy of the quality profession. Let’s look at a few quality dimensions. Quality features are the value add in a smart phone. Perceived quality is the value add to branding. Conformance is the basis for ISO 9001 and other management system standards.
You get the idea. In one article, Garvin structured quality into a profession. These dimensions resulted in quality moving out of the statistical focus of Deming into a strategic business focus. Companies could now differentiate themselves using these quality dimensions and compete on quality. What’s interesting is each of these quality dimensions can be defined, measured, and quality assured for trust.
AI Dimensions
AI is similar but a little different. AI dimensions include: ethics, explainability, performance, privacy, fairness, risk, auditability, and trust. You’ll notice that many AI dimensions are similar to Garvin’s quality dimensions.
Let’s just take one important dimension: trust. Trust is woven into many AI dimensions including ethics, explainability, fairness, and privacy. The challenge is AI trust is extraordinarily difficult to manage and assure. AI is now notorious for its ability to cheat, deceive, and hallucinate—all of which destroy trust.
No Trust – No ROI
According to Edelman Trust barometer, only 32% of Americans trust AI. Why is AI trust such a big deal? Trillions of dollars have been invested in AI. If there’s little trust, then folks won’t use AI to its fullest. And, no AI trust leads to no ROI. And, then the AI bubble goes bust.
So, what’s going to happen in 2026? Forrester Research predicts in 2026 AI trust is going to be the #1 organizational challenge and recommends:
“With change unfolding so quickly, prioritize building trust and help buyers realize value. Invest in AI governance, elevate expert insight, and prepare your teams to complete in a market where buyers demand hard proof.”[i]
Bad and Bad Trust
Trust has two states: bad trust and good trust. We have all experienced bad trust with products, services, and now AI. Product recalls = Bad trust. Customer complaints = Bad trust. What else comes to mind?
The same bad trust applies to AI. Inaccurate credit score = Bad trust. Denial of a house loan = Bad trust. Inaccurate product recommendations = Bad trust. What else can you think of?
We all want good trust with our products and services that are AI generated. As quality professionals, we know the ‘trust but verify’ signals quality assurance. ISO 9001 third party certification is a trust signal. Quality branding, product inspection and internal auditing are assurance signals of trust.
Trust: Heart of AI
Check out this AI trust scenario. An agent making autonomous decisions reveals the difficulty of developing AI trust. First the inputs into the AI system. How do we know the AI system was trained on good data? Next, the AI process that makes the decision. How do we know how the black box algorithm works? Outputs: can we draw a causal or even correlative relationship between inputs and outputs. These are the essence of an AI system audit trail. And, each step of the audit trail has trust challenges. There are a few lessons to be learned from the above story. AI is a relatively new discipline. Many AI attributes are difficult to define and even measure.
Quality has to be built into a product. Quality cannot be inspected into the product after the fact. In the same way, trust has to be built into AI systems through trust frameworks.
Trust Frameworks
Quality has trust frameworks, standards, and best practices. In quality, total quality management is a trust framework. ISO 9001 is a world renowned quality trust standard. Six sigma and lean are operational excellence practices.
ISO 42001 is a new quality trust framework. ISO 42001 is an AI management system (AIMS) that was published in December 2023. ISO 42001 is the first international management system standard for AI. What’s important is that AI is now auditable, and certifiable by third party Certification Bodies. So, if a CB in conducts an ISO 42001 audit, the trust should be the same across national borders.
Your Call to Action
Designing trust into a product or service is a business opportunity for quality professionals. In a previous Quality article, we discussed ‘Show me the Money’. Trust can be an AI and physician jointly diagnosing an illness. Trust can be designed into an autonomous vehicle so it does not run over pedestrians. Trust can be an agent making fair decisions. Quality professionals are key in this area.
Trust is the cornerstone of the quality profession. Quality assurance professionals can be in the driver’s seat for building a new taxonomy built around AI trust, auditing, and assurance.
The science and engineering of AI trust are nascent. They are being developed. This is a huge area for quality professionals to excel, generate revenue, and enhance personal branding. Are you ready to be an AI trust engineer?