Decoding Human-Like Judgments: The Impact on AI Acceptance in Business

As companies race to implement artificial intelligence, they’re uncovering an unforeseen reality: Even the most logical business purchasers aren’t making entirely logical choices — their underlying needs extend well beyond the traditional software assessment criteria.

Let me recount a story: It’s November 2024; I’m situated in a New York City tower, collaborating with a fashion label on their inaugural artificial intelligence assistant. The avatar, Nora, is a 25-year-old digital aide presented on a six-foot-high kiosk. She boasts smooth brown hair, a stylish black outfit, and an engaging smile. She greets clients with a wave when recognizing their faces, nods as they converse, and responds to inquiries regarding corporate history and tech updates. I arrived equipped with a standard technical checklist: response accuracy, conversation speed, facial recognition accuracy…

However, my client didn’t even look at the checklist. Instead, they questioned, “Why doesn’t she possess her own personality? I inquired about her favorite handbag, and she didn’t respond!”

Transforming how we assess technology

It’s remarkable how swiftly we overlook these avatars are not human. While many express concerns about artificial intelligence blurring the lines between humans and machines, I perceive a more immediate dilemma for companies: A foundational transformation in how we assess technology.

When software starts to resemble and behave like a human, users cease to evaluate it merely as a tool and begin to assess it as a person. This occurrence — judging non-human entities based on human criteria — is anthropomorphism, which has been extensively examined in human-pet interactions, and is now surfacing in the human-AI connection.

In the realm of acquiring artificial intelligence products, corporate decisions are not as logical as one might assume since decision-makers are still human. Research indicates that subconscious perceptions influence most human-to-human exchanges, and corporate purchasers are no exception.

Therefore, businesses engaging in an AI contract are not merely entering into a “utility contract” aimed at cost-saving or revenue enhancement; they are entering an implicit “emotional contract.” Frequently, they may not even recognize it themselves.

Perfecting the ‘AI baby’?

Although every software offering has historically included an emotional aspect, once the product starts to closely imitate a real human, this dimension becomes significantly more noticeable and instinctive.

These subconscious responses shape how your employees and clients interact with AI, and my experiences highlight how widespread these reactions are — they are undeniably human. Reflect on these four instances along with their underlying psychological concepts:

When my client in New York inquired about Nora’s favorite handbag, seeking her personality, they were engaging with social presence theory, treating the artificial intelligence as a social entity that needs to feel present and authentic.

One client fixated on their avatar’s grin: “The mouth reveals too many teeth — it’s disturbing.” This response indicates the uncanny valley effect, where features that are nearly human-like create discomfort.

On the other hand, a visually attractive yet less functional AI agent garnered accolades due to the aesthetic-usability effect — the notion that allure can surpass performance shortcomings.

Yet another client, a fastidious business proprietor, continuously postponed the project launch. “We need to make our AI baby flawless,” he asserted in every meeting. “It must be perfect before we can unveil it to the public.” This fixation on constructing an idealized AI entity implies a projection of an ideal self onto our AI creations, as if we’re engineering a digital being that embodies our highest ambitions and standards.

What is most significant for your business?

So, how can you lead the market by leveraging these concealed emotional contracts and outpace your competitors who are merely accumulating one advanced artificial intelligence solution after another?

The essential element is identifying what is crucial for your business’s distinct needs. Establish a testing protocol. This will assist you not only in pinpointing top priorities but, more critically, in deprioritizing minor details, regardless of how emotionally engaging they may be. Given that the sector is quite new, there are virtually no readily available manuals. However, you can be the trailblazer by creating your unique method for understanding what aligns best with your business.

For instance, the client’s inquiry about “the AI avatar’s personality” was validated through testing with internal users. Conversely, most individuals were unable to discern differences between the multiple versions the business owner had vacillated over for his “perfect AI baby,” indicating that we could settle at a “good enough” juncture.

To assist you in recognizing patterns more effortlessly, consider bringing on team members or consultants with a background in psychology. All four examples are not isolated instances but rather well-researched psychological phenomena occurring in human-to-human interactions.

Your relationship with the technology vendor must evolve as well. They should become a partner who navigates the experience alongside you. You can arrange weekly meetings with them post-contract signing and share your insights from testing so they can produce improved products for you. If budget constraints exist, at least allow extra time for product comparison and user testing, revealing those hidden “emotional contracts.”

We are at the forefront of defining how humans and artificial intelligence engage. Successful business leaders will embrace the emotional contract and implement processes to navigate the uncertainty that will enable them to seize the market.

Joy Liu has steered enterprise products at AI startups and cloud and artificial intelligence initiatives at Microsoft.