Lenka Hámošová: We need to Increase AI literacy. We cannot avoid manipulation.

How do we recognize original design in a world saturated by AI? Lenka Hámošová, a respected visual artist and researcher, reflects on current trends in design where artificial intelligence plays an increasingly significant role. Specializing in synthetic media and AI in art, Lenka discusses the challenges of identifying original work in the AI era and offers a perspective on how to navigate this rapidly evolving environment.

Category:

Speculative Design

Author:

Pavel Danyi

Read:

11 mins

Location:

Prague

Date:

July 21, 2025

Lenka, what led you to work with synthetic media and artificial intelligence?

My professional background is in critical design. I have always been involved in projects exploring various social phenomena that leave people feeling disoriented or misunderstood. At the time, I primarily focused on visual communication and transparency—issues like "visual smog" in cities or "information bubbles" on the internet. I explored how our constant online presence affects our ability to perceive information visually and how we become susceptible to manipulation.

Then, in 2018, deepfakes emerged, which was a direct continuation of the visual transparency issue. I realized people would struggle to understand and correctly evaluate what is reality and what is fiction. This led me to dive deeper. I found that striving for absolute visual transparency is a hopeless and futile task. Instead, it became crucial to inform the public about the existence of deepfake tools.

Creating synthetic content has a much broader scope, which is why it's important not just to study these technologies but to apply them in practice. I started organizing workshops and designing tools to help people better navigate what can be "faked" using AI. That was the foundation of my focus on this field.

You mentioned "tools"—I noticed you use special cards in your workshops. How do they work?

Originally, the workshop was designed specifically around these cards, but I eventually integrated them into all my sessions. Organizing them into color-coded categories helps participants understand the possibilities of AI and machine learning more quickly. It allows them to distinguish between models for video, image, text, and sound, or models used for post-production.

In the last year and a half, the shift toward multimodal models has made traditional categorisation less precise.

However, the cards remain a valuable resource, containing specific model names and links to academic papers or GitHub pages for open-source projects. I am currently working on a complete redesign of the card system to adapt to current AI trends. It’s a fun design challenge—like a logic puzzle to give it a structured flow.

3. The Good Time Journal

The third tool was the Good Time Journal. This method goes deeper than the dashboard, tracking daily activities based on two parameters: Engagement (Flow) and Energy (does this activity drain me or fill me up?).

I kept this log for three weeks.

The Findings: I noticed that creative activities—like morning sketching or finishing a research segment—put me into a state of flow and energised me. Conversely, reactive tasks like answering emails or fixing organizational snags drained me with zero added value.

The Pattern: My days lacked the variety I craved, often slipping into stereotypes. However, visualising this allowed me to redesign my weeks. I began to consciously schedule "recharging" activities and limit the energy vampires

How have technologies like deepfakes, often seen as threats, influenced your professional life and worldview?

Deepfakes don't scare me anymore because I realized that fighting against them is futile. It is concerning that our parents, and perhaps even my generation, cannot reliably recognize a well-executed deepfake—but that is their purpose. A few years ago, you could spot imperfections like blurred pixels or inconsistencies. Today, generative AI tools are becoming so realistic that they are almost indistinguishable from reality.

That is why AI literacy is key. We must teach people that even a recording can be manipulated. History shows that every new technology brings a period of confusion—just like when photography appeared and people used it to create seemingly supernatural phenomena.

Today, I am increasingly skeptical. I don't spend much time on minor news, but for key information, I take the time to verify. I don’t trust a single video or source; I look for confirmation elsewhere. A bigger problem today is people labeling true news as "fake news" or "deepfakes" just to dismiss it. This has made me much more critical of the information I consume.

How has the design world changed due to AI?

I’ve noticed many people using AI to simplify their work, which is great. Tools like Generative Fill in Photoshop or text-to-vector in Illustrator significantly increase efficiency. When combined with ChatGPT for brainstorming, productivity skyrockets.

However, I see designers relying heavily on Midjourney visuals. This raises the question of how much these designs are further edited or if they are just considered a final product. Relying solely on Midjourney feels insufficient to me; every good design needs a well-thought-out creative process.

Experienced designers will likely be more cautious with AI integration to avoid poor quality. On the other hand, generated vectors are very useful for generic tasks like stock imagery or standard corporate graphics. If someone was previously creating generic vectors and now AI does it, the problem lies in the generic nature of the work itself, not the tool.

There are also new opportunities for UX and UI designers. Users are looking for simple, intuitive interfaces for AI, unlike the complex Discord servers or Google Colab Notebooks used now. This evolution raises questions about user freedom and transparency—interfaces should not hinder interaction but should remain transparent about the models and datasets being used.

What are the main ethical dilemmas you encounter?

The biggest issue is unfair datasets. They lack transparency. They often contain licensed content or project human stereotypes through image descriptions. Ideally, institutions like libraries or universities should create their own datasets to solve this.

Another problem is that private companies' rules may not align with ethical standards.

We are using technologies we don't fully understand, which can have unknown impacts. In creativity, this might not seem dramatic, but when AI is used in hiring or healthcare, the ethical dilemma becomes vital. There is also the unresolved issue of copyright and attribution—how much was human and how much was machine? This is something academia must address as well.

Would you like me to summarize the key takeaways of this article for a presentation or a social media post?

What is Speculative Design? (From an IKEA Catalog to a Marketplace Ad)

The brief wasn't simply "create an illustration." My task was to apply the method of speculative design. The goal of this method isn't to predict the future, but to materialise possible scenarios – often controversial ones – so we can discuss them before they become reality.

In the course of our collaboration, I generated dozens of different variants and artifacts, from wearables to digital services. For the purpose of this article, however, I will focus on one specific concept that elicited the strongest reactions – the PulseNest.

Roman Sellner and Natálie Káčová provided me with a scenario of the world in 2050, where technology in the care of grieving parents is a common commodity. My task was to create a believable ecosystem for this world. It wasn't enough to just create a product. I had to show its entire life cycle: how it is sold, how it is reported in the news, and what happens to it when no one wants it anymore.

The power of speculative design lies in its banality. If the artifacts looked like props from Black Mirror, the viewer could easily dismiss them as unrealistic. But if they look like something that pops up in your morning news feed or on Facebook, the defense mechanisms crumble, and a sense of chilling unease sets in.

Phase 1: The Commodification of Grief (The Catalog)

The first step was to create a product that promises comfort. I placed the PulseNest artifact within the aesthetic of a department store catalog.

Fig. 1: PulseNest in the aesthetic of a global retail chain. The visual combines a homely atmosphere with a technological description of a product that "plays back the actual heart rhythm and breath of your child." The £990 price tag adds a terrifying reality – grief becomes an item in a shopping cart.

Phase 2: Institutionalization (News 2050)

For the world to be believable, society must react to technology. Therefore, I created a fictional article from the iDNES.cz news server dated 2050.

Fig. 2: Fictional article "PulseNest Enters Czech Hospitals." The text simulates public debate – quoting enthusiastic ministry spokespeople as well as skeptical psychologists warning against "pathological mourning." This artifact demonstrates how such technology would be normalized in the public space.

Phase 3: Waste and Forgetting (Marketplace)

The most chilling part of my work was the visual showing the end of the product's life cycle. What happens to a pillow that simulates the heartbeat of a dead child when the parents no longer want it?

Fig. 3: Ad on Facebook Marketplace. The text "Selling our PulseNest, we no longer need it" for 5,000 CZK is a brutal collision of sacred grief and banal consumerism. It shows the moment when a "tool of love" becomes "used goods" that need to be discarded.

Methodology: battling AI and finding "The Uncanny"

Although the result may look simple, the path to it was long. My main tool was generative artificial intelligence (Perplexity for context, ChatGPT/DALL-E 3 and Gemini for visuals), but it wasn't a "one-click" job.

It was more of a battle with the algorithm. I generated hundreds of variants and variations, testing different styles and compositions. AI tends to generate "pretty" and "happy" people from stock photos. My task was to force it to generate an emotion that is much more complex: a mixture of hope, pain, and shame.

Case study: The evolution of a prompt

What does the search for the right expression look like for a mother using such a product? Here is an example of how I had to iterate the assignment (prompt) to get from kitsch to something that works.

Attempt No. 1 (Failure): Prompt: "Sad mother looking at a computer screen with an avatar of a child, futuristic style, blue light." Result: Too theatrical, too "cinematic." The character was looking directly at the viewer, which felt like a cry for help. It lacked intimacy and believability.

Attempt No. X (Dead End): Prompt: "Product photography, realistic lifestyle shot. A woman sitting at a kitchen table. Looking at a tablet. Warm lighting." Result: Too cozy. It looked like an advertisement for video calls. The uncomfortable ethical contradiction was missing.

Attempt No. 42 (Final Version): Prompt: "Cinematic shot, hyper-realistic. Over-the-shoulder view of a tired woman looking at a tablet screen. The woman is NOT looking at the camera; she is avoiding eye contact, looking down, expression is a mix of comfort and deep shame. Soft, cold morning light. Commercial aesthetic but with a dystopian undertone." Result: Only here did I manage to hit the key emotion – shame. The woman avoids eye contact because she knows what she is doing is socially unacceptable, but she can't help herself. It was this detail, fought for through dozens of attempts, that gave the artifact its believability.

Impact: When Design Causes Terror

Since I was working remotely, I had no direct contact with the workshop participants (healthcare professionals and experts). I often doubted whether mere pixels sent over the internet carried any weight. The true impact of this method only hit me when listening to the podcast Ministerský jednorožec (The Ministerial Unicorn), where psychotherapist Kateřina Hájková Klíčová reflected on the participants' reactions to these very visuals.

The artifacts worked exactly according to the definition of speculative design – they induced an "ethical shock." Participants did not react to the aesthetics, but to the terrifying availability of these technologies. As stated in the podcast:

"I see terror in your eyes. Speculative design can do that. [...] Only when we go through that shock can we then ask: What will we, as humans, do about it?" [^1]

When the healthcare professionals saw the catalog with the PulseNest pillow, their first reaction was rejection: "We don't want this, let's go back to bonfires." And that was exactly the goal. Thanks to these artifacts, they were able to experience a "dry run" of a situation where a parent comes to them requesting such technology. We created a safe simulator for dilemmas that await us.

© Visual Journal ジャーナル

(WDX® — 02)