Our Digital Doppelgängers: Synthetic Humans

“More human than human.”

This quote comes from Eldon Tyrell, the founder of the Tyrell Corporation, which manufactures the synthetic humans, called “Replicants,” in the sci-fi cult classic Blade Runner. It is the motto of his company and directly conveys how advanced and human-like the Replicants in this world already are. However, the quote also makes clear that the boundaries between human and machine have already been completely blurred. The synthetic humans have been created so perfectly that they appear more human than actual humans in many ways. A distinction is no longer possible.

This quickly leads to the central question of the film: What makes a human a human? If a synthetic human, if AI, manages to simulate emotions perfectly, can remember, and is externally indistinguishable from real humans, then where does the difference lie in perception from the outside? This question runs throughout the entire film and challenges viewers to question their own ideas of humanity.

But what does a sci-fi film have to do with the AI logbook? Today, we have already approached an important threshold that makes films like Blade Runner appear at least more realistic than they did a few years ago. Because drum roll, they already exist today - synthetic humans. For now, they only live as software constructs in our computers, but the step into the real world is not a big one, thanks to simulations.

Synthetic Humans - The Next Revolution in User Research?

What could synthetic humans be used for, you ask? For example, in the research of user behavior. Sounds crazy? This is precisely what new AI technologies such as “Synthetic Humans” by Fantasy Interactive or the “neuro-symbolic AI” by Lakmoos promise.

These systems generate artificial users meant to simulate human behavior, opinions, and preferences. The AI agents can wake up, have breakfast, and go to work. They form opinions, notice each other, and start conversations. They remember past days and plan the next. Stanford University, together with Google, published an exciting study on this topic last year titled “Generative Agents: Interactive Simulacra of Human Behavior.”

The advantages are clear: lightning-fast results, unlimited sample sizes, and significantly lower costs compared to traditional surveys. Lakmoos even promises 95-98% accuracy in predicting user behavior.

But can synthetic data really reflect human behavior in all its complexity?

Opportunities and Limitations of AI-Supported Market Research

The possibilities are fascinating: product developers could receive feedback on new features in seconds. Marketing experts could test various campaign ideas on thousands of virtual people. City planners could simulate the effects of new traffic systems.

At the same time, there are legitimate doubts about the validity of such data. Critics argue that synthetic humans cannot authentically replicate the subtlety of human emotions and irrational behavior. Furthermore, there is a risk that existing biases and distortions in the training data could be reinforced.

A balanced approach seems sensible: AI agents as a valuable addition to traditional research methods, not a complete replacement. Human expertise in interpreting the data and uncovering deeper motivations remains indispensable.

Ethical Questions and Social Impacts

The development of synthetic humans also raises important ethical questions:

• How do we ensure that simulated users represent the diversity of society?
• Who is liable if AI agents exhibit discriminatory or harmful behavior?
• How transparent must companies be in disclosing the use of synthetic data?

Additionally, we must consider the long-term impacts on the job market. Will thousands of jobs in market research disappear? Or will new professions emerge at the intersection of AI and human expertise?

Tips for Handling Synthetic User Data

1. Try combining synthetic and real data for a balanced view.
2. Keep a sharp eye! Critically question the training data and algorithms of AI systems.
3. Use the efficiency of AI to gain more time for in-depth qualitative research.
4. Be transparent with customers and users about the use of synthetic data.
5. Invest in training to competently interpret AI results.

In conclusion, I’d like to leave you with a quote from the philosopher Immanuel Kant:

“The science of man is the most important of all sciences.”

Let’s make sure that even in times of AI and synthetic humans, humanity remains at the center.