What an Egyptian priestess taught us about AI-generated Synthetic Users.

Photo by British Library on Unsplash

The year is 530 B.C.

Egypt is about to be conquered by the Persian empire.

As an Egyptian priestess, you begin your day with a purification ritual, cleansing yourself in the sacred waters of the Nile. Once you have adorned yourself as the embodiment of the divine you are, you head off to the temple. There, you prepare a basket of fruits, bread, and wine for Ra, the Sun God, seeking his radiant energy to bless your kingdom. Mid-day, you help the statue of Anubis, the guardian of the afterlife, in its daily routine of cleaning and anointing its dark stone with precious oils. You know how cocky he is and how much he likes to look pristine. Plus, today, you have scheduled a few funeral rituals and want him to be particularly happy to safely guide souls to the realm of the dead.

Before calling it a day, you confide in the statue of gracious Isis, goddess of magic, your fears about the Persian invasions. She listens to you and, in response, subtly moves her stone head. She has spoken: that’s the sign for you to sacrifice a young, pure, healthy cow in her honour.

Your day is over. It ends with a sense of spiritual fulfilment as you lay in your reed mat. Little do you know that you will become a UX researcher in your eleventh reincarnation.

The year is 2027 C.E.

Things have changed — it’s not that easy to find pure and healthy cows nowadays — but the human fascination for mirroring human traits in our creations and interacting with them as they have a mind of their own remains as vivid as ever. Whether born from chisel and stone or lines of code, they still empower us, offering guidance and augmenting our capabilities. And, just as in the past, we approach these marvels with a blend of excitement and fear.

Think about the latest Large Language Models and their capacity to simulate human speech and behaviour. They are revolutionising industries and jobs from content generation to healthcare diagnosis. While a nod from Isis served you well at the time, you feel very fortunate that, in this life, you can count on a bit more sophisticated innovations to conduct critical tasks such as User Research.

How? Working with Synthetic Users, virtual characters created to represent real users in user research.

Among the multiple advantages they have to offer, these stand out:

1. Time-to-market: they can be created much faster than traditional personas, which can take weeks or even months to develop. This makes them ideal for situations where quick turnaround times are required, such as in agile software development environments.
2. Adaptation: they can be easily modified and updated, allowing researchers to adapt to changing user needs and preferences.
3. Controlled environment: you have complete control over the testing environment, allowing you to isolate variables and minimise external factors that might impact your tests.

Anatomy of a Synthetic User: First layer, facts and evidence (Real data from organic users who create the identity structure of the US); Second layer, assumptions and inference (Informed assumptions AI is able to make based on traditional users’ information and real responses); Third layer, imagination and transcendence (Ability to go beyond programmed or learned knowledge and generate creative and novel ideas)

While there is no doubt about the immense usefulness of Synthetic Users in reshaping the landscape of user research and design, it’s crucial to acknowledge that this technology is still in its infancy. In this uncharted territory, adherence to a set of guiding rules becomes paramount. As a former Egyptian priestess, you remember the Ma’at, the sacred Egyptian rules that guided your life and ceremonies two millennia ago. Based on these principles, we’ve created for you the Ma’at version for Synthetic Users:

Ma’at Sacred Principles vs.

Synthetic Users’ Sacred Principle

1. Truth

Be clear about what organs go in which canopic jars. The gods need to understand the input to grant the afterlife.

Transparency

Be clear about how the SUs are defined and what data they use to make decisions. We need to understand the input to interpret the output.

2. Justice

Ensure your biases and personal preferences do not influence which goat you sacrifice to Ra and which to Isis.

Bias

Ensure data is diverse and representative to avoid biases. Plan regular audits using counterfactual and adversarial testing.

3. Protection of the Vulnerable

Make sure those in need are protected and safeguarded, providing them with food, clothing, and shelter.

Privacy

Make sure users are aware of what data is collected, how it will be used, and who has access to it. Ask for consent to comply with regulations (GDPR, CCPA, etc.).

4. Balance

Understand the Nile’s needs both inundation and the drought periods, to optimise its performance.

Human Oversight

Understand SUs need human checks and balances to optimise their performance. Detect inconsistencies and errors. Evaluation methods such as ROUGE could be helpful.

5. Harmony with Nature

Spirit, humans, animals, plants, etc., co-exist in ways that create alignment and beauty. Each expression of being must be authentic. It is only through authenticity that harmony can truly be achieved.

Continuous improvement

Human oversight must ensure authenticity of the SU. Detect inconsistencies and errors. Evaluation methods such as ROUGE could be helpful.

These principles ensure the ethical and effective use of SUs and pave the way for responsible innovation. As we navigate this exciting frontier, let’s remember that our actions today will shape the future of user-centric design, and it is our collective duty to do so with wisdom and integrity.

Ready to harness the power of Synthetic Users in your research projects? Reach out to us for personalised guidance.

Let’s take your user research to the next level together!