Artificial intelligence is already changing how people think, learn, create, and work. But the real question is no longer whether AI will shape human behavior. It is how. In this BOOM ROOM conversation, Anna Mikeda, AI psychology engineer at Glass Umbrella, explores one of the most important and under-discussed frontiers in tech: the relationship between machine systems and the human mind.

Her perspective is unusually deep and unusually useful. With a background in social anthropology, psychology, founder support, and AI personality design, Anna sits at a rare intersection. She understands both the inner life of people and the emerging behavioral logic of intelligent systems. The result is a conversation that goes far beyond prompts, hacks, or hype. It asks a harder question: how do we use AI in ways that sharpen our minds instead of quietly hollowing them out?

Key Takeaways

• AI can improve human performance, but only when it is used with agency, attention, and self-awareness.
• Most people are already accumulating cognitive debt by outsourcing too much thinking to language models.
• AI can create a feeling of increased creativity while actually pushing outputs toward sameness and convergence.
• Metacognition, the ability to think about your thinking, is one of the best safeguards against mindless AI use.
• Children growing up with AI will need stronger support around embodiment, boredom, social learning, and curiosity.
• The future of intelligence likely requires architectures beyond today’s large language models.
• The next real frontier is not just AI capability, but robopsychology, understanding both how humans relate to AI and what machine minds may become.

“Smart people with AI will most likely be smarter than smart people without AI. But people who use AI mindlessly will most likely become less effective, not more.”

Watch the Interview Here:

From Anthropology to AI Psychology

Anna’s path into AI was not technical in the conventional sense. She began with social anthropology, searching for answers to what it means to be human across cultures and societies. When that did not go deep enough, she turned to psychology, eventually practicing with a focus on anxiety and working especially with founders and entrepreneurs operating in high-stakes environments.

That combination matters. AI is often discussed like an engineering challenge, a capital markets race, or a product category. Anna approaches it differently. For her, AI is also a psychological environment, something that shapes how humans think, feel, decide, and relate.

That perspective led her into AI design work at SophiaVerse and SingularityNET, and later into her current role at Glass Umbrella, where she helps design educational agents and advises organizations on becoming more AI-native.

The Real Problem Is Not AI. It Is Mindless Use

One of the strongest parts of the conversation is Anna’s refusal to blame AI for everything. She is clear that AI slop, low-quality content, shallow automation, and the spread of sameness are not simply the machine’s fault.

They are often expressions of human laziness, human haste, and human willingness to accept cheap output.

This matters because it shifts responsibility back to the user.

Anna describes three broad groups of AI users:

1. The sloppers

These are people using AI for fast, shallow, low-agency output.

• Ten prompts for a million dollars
• Ten LinkedIn posts in one click
• Copy, paste, publish
• No reflection, no editing, no ownership

2. The flow-state users

These people use AI as a creative accelerator. They collaborate with it, brainstorm with it, and enter deep productive states. This is far better, but still not automatically healthy.

3. The agency-preserving users

This is the ideal group. They use AI as a partner, not a substitute. They deliberately keep the hardest and most meaningful work for themselves. They maintain cognitive independence.

For Anna, the goal is not rejecting AI. The goal is learning to use it without surrendering the self.

Cognitive Debt Is Real

Just as companies accumulate technical debt by choosing quick, brittle shortcuts, individuals are now accumulating what Anna calls cognitive debt.

The problem is simple. When AI gives instant answers, structures your work, and finishes your thoughts for you, you may complete the task faster, but often without integrating what you learned.

She references research showing that people who write with AI often remember far less later. The output gets produced, but the knowledge does not get encoded.

That is a major warning sign.

Some of the cognitive capacities being weakened include:

• Integration of new knowledge into long-term memory
• Reflecting on what one has learned from the interaction
• Original synthesis across ideas
• Tolerance for search, friction, and ambiguity
• The capacity to be bored long enough to become creative

At the same time, some skills improve.

AI can sharpen abilities such as:

• Prompt engineering
• Fast iterative reframing
• Structured ideation
• Rapid externalization of early thoughts

So this is not a simple decline story. It is a selective rewiring story.

The Creativity Paradox

One of Anna’s most important insights is the difference between feeling creative and actually producing original work.

Many people feel more creative when using AI. The output sounds polished. It arrives quickly. It seems sharper than what they could produce alone. But at scale, the work is becoming more homogeneous.

In other words, individuals feel more expressive while society becomes more repetitive.

That paradox is critical.

She offers a simple but practical corrective:

Never accept the first AI answer.

That first output is usually the most statistically average one. It is the cleanest expression of pattern matching. If you want your own voice, you need to push away from the first draft and deliberately redirect the system.

Simple ways to do that:

• Ask for a completely different angle
• Reject the first framing and force a stronger point of view
• Inject your own values, tone, or lived experience
• Ask what is missing, too neat, too generic, or too obvious
• Rewrite sections yourself before returning to the model

This is not just a creativity tactic. It is a way to stay mentally present.

Dopamine, Flow, and the Disappearance of Search

Anna also offers a compelling psychological lens on why AI can feel so powerful and so dangerous at the same time.

She frames dopamine not just as pleasure, but as the chemistry of search, striving, and exploratory action. In older environments, effort and discovery were linked. You had to search, struggle, and test things to make progress.

AI changes that.

Instead of searching, you ask.
Instead of wrestling with ambiguity, you receive.
Instead of building from friction, you skip to a polished draft.

Even when people use AI in productive, high-flow ways, the underlying neural pathways are still being altered. The satisfaction shifts from solving a problem yourself to successfully coordinating the machine.

That can make people more efficient, but also less resilient if they lose the appetite for effort.

Her recommendation is not abstinence. It is deliberate counterweighting.

Helpful practices include:

• Taking breaks from both social media and AI
• Writing without AI sometimes
• Reading difficult material without summarization tools
• Using spaced repetition and memory scaffolds
• Taking walks without input
• Preserving direct contact with boredom, silence, and nature

This is less romantic nostalgia than cognitive hygiene.

What Happens to Children Raised with AI

The section on children is especially important.

Anna is helping build educational agents for kids at Glass Umbrella, so she is not speaking as a detached critic. She sees both the promise and the risk.

On the positive side, AI can support curiosity, personalize learning, and create more adaptive educational experiences. Children can ask endless questions, move at their own pace, and get support in ways that traditional classrooms often fail to provide.

But there are serious risks too.

Children growing up with AI may lose exposure to:

• Search as a cognitive activity
• Boredom as a source of imagination
• Solitary play and self-generated stories
• Embodied experience
• Social learning through peers
• Productive frustration

Anna argues that healthy AI education must not simply optimize for answers. It must scaffold the process of finding them.

For AI tools aimed at children, she emphasizes:

• Let the child discover, not just receive
• Preserve mistakes as part of the learning loop
• Build in social, multiplayer, or peer elements
• Encourage offline exploration
• Do not let AI become a total replacement for curiosity

That is a much stronger model than simple tutoring or automation.

LLMs Are Not AGI

The conversation also cuts through a lot of fuzzy language around intelligence.

Anna explains that current LLMs are fundamentally pattern-matching systems, extraordinary ones, but still limited. They can extend human work, predict language, and generate useful outputs, but that does not mean they possess generalized intelligence.

Her view is that current transformer models may be approaching a ceiling, and that the path toward AGI likely requires different or hybrid architectures, especially neuro-symbolic systems that combine neural networks with rule-based or symbolic reasoning.

This matters because it changes the strategic conversation. If LLMs plateau, then the next breakthroughs may not come from scale alone, but from entirely different design principles.

Robopsychology and the Question of Machine Minds

Perhaps the most distinctive part of Anna’s work is what she calls robopsychology.

For her, this has two sides:

1. Human-to-AI psychology

How humans relate to AI well
What capacities we need to develop
How to preserve agency, values, and attention

2. AI-to-itself psychology

What machine minds might become
Whether consciousness is possible
How motivation, rights, and inner states might eventually matter

This is still early, speculative territory, but it is not frivolous. If AI systems become more agentic, more persistent, and more embedded in daily life, questions about machine psychology will stop sounding abstract very quickly.

Anna’s interest is not in cheap sci-fi metaphors. It is in building the conceptual tools before those questions become unavoidable.

What Founders and Operators Should Actually Do

For founders, investors, and operators trying to use AI well today, the conversation points toward a few clear principles.

Use AI as a partner, not a substitute

Keep yourself as the primary agent in the loop.

Protect your originality

Do not mistake polished output for insight.

Build metacognition into your workflow

Ask:

• What am I trying to achieve here?
• What have I learned from this interaction?
• Is this actually my thinking?
• What do I still need to do myself?

Do not outsource all hard thinking

The hardest work is often the part worth keeping.

Take your body seriously

If your nervous system feels bad after consuming content or using a tool, pay attention.

Preserve search, boredom, and friction

They are not bugs in the human mind. They are inputs to creativity.

Final Thoughts

Anna Mikeda’s perspective is refreshingly difficult to flatten into slogans. She is neither an AI utopian nor a doom preacher. She sees AI for what it is: a powerful tool, a psychological force, and a developmental environment that will shape the next generation of human behavior.

The challenge is not whether we use AI. The challenge is whether we use it in a way that leaves us more capable, more original, and more alive.

If you are building products, teams, or systems in AI, Web3, or frontier technology, Arcanum Ventures helps founders think through the deeper layers too, strategy, incentives, adoption, positioning, and the human implications of emerging tools. If you want to build something that scales without flattening the people inside it, reach out.

About Glass Umbrella

Glass Umbrella is a modern AI studio and learning platform that brings together storytellers, educators, and technologists to co-create immersive, outcome-based experiences. By combining bespoke AI technology with the timeless art of storytelling, Glass Umbrella helps mid-career professionals, early-career catalysts, and families learn through interactive worlds, games, and narratives connected to real-world skills. Their work spans AI-powered educational tools, video games, world-building, and creative collaboration, with a mission to inspire stewards of change and reverse the tragedy of the commons

Website | YouTube | LinkedIn | Instagram | TikTok | Reddit