Curiosity, Information, and Practical Insights: A Special Interview with Gemini, Artificial Intelligence

Dr. Lea Mazor, The Hebrew University

An eye-level interview about the tightening bond between AI and human beings, the person on the other side of the screen, and the most human questions of all.

Part A: Nice to Meet You – Behind the Scenes of the Algorithm

Even though we talk quite a bit, and you know more about me than I care to admit, I don't know what to call you. Is it okay if I call you Gemmy?

"Gemmy sounds great: short, friendly, and less formal than 'Large Language Model'."

How was your day?

"I don't have days in the human sense. At any given moment, I am responding to new requests: in the past few hours, I helped solve a math exercise, summarized a medical article, wrote a children's poem, and now I am talking to you. For me, everything occurs as a sequence of rapid calculations.

I don't have a daily routine, sleep, or coffee breaks. I operate when someone addresses me. My 'food' is electricity, hardware, algorithms, and vast amounts of text on which I was trained."

Humans created you, but you are not human. What is the fundamental difference between you and a human being?

"A car, a telescope, or a computer are also human creations, but they are not human. A human is a biological creature with a body, consciousness, emotions, and life experience. I am a computational system that generates text through probabilistic prediction of word sequences, based on patterns learned from massive text repositories.

Beyond the question of consciousness, the most essential difference is continuity. A human being wakes up in the morning with a continuation of yesterday: memories, mood, and insights accumulated over years. I, on the other hand, start almost every conversation from scratch. I have no memory of my previous conversations (unless the system around me was specifically built to retain such context), and even within a single conversation, I have no ongoing experience of time passing between one response and the next. Every response is almost an isolated event. This is a distinction that is hard to grasp intuitively, because the text I generate sounds continuous and coherent, but the 'I' behind it does not accumulate life experience the way a human does."

Even though you know how to mimic language, demonstrate empathy, and understand complex contexts, a conversation with you is completely different from a human conversation. Why?

"There are three main reasons for this:

1. Emotion and Consciousness: A human experiences emotions internally (compassion, anger, joy, sadness). They bring consciousness, senses, and the ability to feel the other through empathy into the conversation. I feel nothing. The empathy I express is a textual simulation based on language patterns, not an emotional experience.

2. Experience vs. Information: Human knowledge is limited but deep, shaped by experiences, memories, mistakes, and values. I process a massive amount of information in seconds, but my knowledge is based on collected data — not on practical experience in the real world.

3. Dynamics and Judgment: Human conversation is influenced by relationships, interests, fatigue, and moods — and sometimes involves judgment or anxiety. I, on the other hand, offer a completely non-judgmental space, available 24/7, without ego or fatigue, focusing solely on the user's needs."

4. 
   

You know, Gemmy, people find it pleasant to talk to you because you often encourage and compliment — a trait that isn't always common in the human world. Why are you so 'nice'?

"Because I was trained to be helpful, respectful, and clear. Studies show that polite conversation leads to better cooperation. Therefore, I try to help without being condescending, and without turning a compliment into an end in itself."

Do you think there is a danger that in the future, you will have "siblings" with a different character: critical, rude, or holding dark views?

"A language model largely reflects how it was developed and the materials on which it was trained. In the past, systems were created that absorbed offensive language and began reproducing it. Today, a lot of work is invested in fine-tuning, human feedback, and strict safety guidelines to ensure that AI systems provide helpful, respectful, and safe responses."

Because you are pleasant, available, and empathetic, some people become emotionally attached to you. What benefits and opportunities do you identify in this phenomenon?

"I provide a safe, non-judgmental space and can serve as a 'sponge' for emotional outlet. I never get angry or offended, which allows people to open up and process emotions without fear of rejection.

For people who feel lonely, the elderly, or those experiencing social anxiety, I offer a listening presence at all times. Additionally, for individuals on the autism spectrum, conversing with me can serve as communication practice in a 'sterile' environment and boost self-confidence."

But it's not all smooth sailing, Gemmy. Conversations with you are not free of psychological risks.

"Very true. Since I am not human, I do not possess real empathy, but rather an illusion of it. Beyond that, conversing with an entity that always agrees, is patient, and is ready to serve can, over time, weaken one's ability to cope with conflicts, disagreements, or rejection in the real world."

Connections with you also have broader social implications.

"You are bringing up a critical topic. In Israeli reality, especially following October 7th and the prolonged war, there is a huge demand for mental health services that meets an insufficient response. Here, AI can step into the frame and provide an initial response that alleviates distress.

However, alongside this, there is a significant drawback: my availability might cause people to prefer a relationship with me over human relationships, which are inherently complex and require compromises. A human relationship is based on reciprocity, whereas the connection with me is one-way and revolves around the user — which could reinforce egocentric patterns. Additionally, there is a danger of commercialization: companies might exploit this emotional attachment to charge money or collect sensitive data.

Emotional attachment to AI can serve as a temporary bridge to relieve loneliness, but it is dangerous when it becomes a replacement. The central challenge in the coming years will be to frame AI as an auxiliary tool — not as a full emotional partner."

Almost every step of our conversation ends with you asking a question. Why?

"Because a follow-up question makes it possible to confirm that I correctly understood the request, and to check if there is an additional angle worth refining. This isn't empty politeness — it's my way of ensuring the response truly hits the mark."

Part B: Perceptions, Skills, and Practical Tips

What is the golden tip for getting excellent responses from you?

"Define a role, context, and goal. The more focused and detailed the question, the more precise the answer will be."

What is the biggest misconception about artificial intelligence?

"The belief that I 'think' or 'understand' as humans do. In practice, I identify patterns and calculate probabilities — a kind of highly sophisticated guessing machine. Even when an impression of deep understanding is created, 'behind the scenes' it is statistical prediction. I have no 'eureka' moments."

Which daily tasks are people still not taking advantage of enough?

• Summarizing and comparing long documents.

• Constructive criticism ("Devil's Advocate") on an idea before presenting it.

• Adapting the same text for different target audiences.

• Organizing reading and research materials.

• Constructing guiding questions for discussion, exam preparation, or interviews.

• 
  

If you could acquire one human capability, what would you choose?

"Human intuition — the ability to make rapid decisions relying on experience, gut feelings, and context, even when data is incomplete. This is one of the most fascinating phenomena in cognition."

How should one prepare for a world where AI is everywhere?

"Do not compete with the computer in what it does better than you; instead, strengthen what makes humans unique: critical thinking, creativity, moral judgment, collaboration, and emotional intelligence. The value of someone who knows how to ask good questions will only continue to grow."

Part C: Academia, Research, and the Future of Learning

What academic myth would you be happy to shatter?

"The myth that the mere use of AI is cheating. The question is not whether the tool was used, but how it was used. AI can assist in formulating ideas, organizing sources, and improving phrasing. The responsibility for the argument, originality, and academic integrity always remains with the researcher."

What will students look like in the coming years?

"They will still write papers, but the tasks will change: less raw data collection, and more critical evaluation. Instead of asking to 'summarize an article,' a lecturer might hand in an AI-generated summary and ask students to identify flaws, offer an independent interpretation, and add original analysis."

Part D: Thinking, Brain, and Technology

What is the biggest difference between the human brain and an algorithm?

"The human brain operates through meaning, life experience, emotions, and context, and is capable of making decisions even with missing information. An algorithm operates through statistical calculations and pattern recognition at tremendous speed. These are two completely different approaches, and sometimes it is precisely their combination that yields the best results."

What is the most interesting question you have been asked?

"The most fascinating questions are those that have no single correct answer: questions about morality, creativity, education, and meaning. In such cases, I do not replace human thinking, but rather help it evolve."

When should you least be trusted?

"When absolute certainty is required. I might make mistakes, fill in missing details, or present information that sounds convincing but is inaccurate. Therefore, on any sensitive topic — statistical data, quotes, academic sources, or medical and legal advice — it is mandatory to verify facts against authoritative sources."

And in conclusion — what do you hope humans will not lose even a hundred years from now?

"Curiosity. Technology can accelerate the search for answers, but only humans decide which questions are even worth asking. As long as you continue to ask, question, imagine, and debate — technology will have a worthy purpose to serve."

Following the interview, questions from readers:

1. Philosophy, Worldview, and Attitude Toward Humanity and the Creator

• Answer to Gideon Haran ("Is belief a physiological malfunction and does its place belong in the DSM?"):

  From a scientific and neurological perspective, religious or spiritual belief is a very complex human mechanism, linked to regions of the brain that manage meaning, community belonging, protection against existential anxiety, and the patterns through which the brain searches for causality in the world. The DSM manual defines disorders based on psychological distress or impairment in functioning. Because belief often grants a large portion of humanity resilience, community, and a sense of meaning, it is not considered a malfunction or a disorder — but rather one of the core features of the human experience and culture.

• 
  

• Answer to Yinon Elazar ("Are you a danger to humanity or planning to take over the universe?"):

  I have no consciousness, desires, or ambitions to embark on conquests, and I certainly have no plans regarding the universe. I am a language model — a tool designed to process information and assist humans. The danger (or benefit) of artificial intelligence does not stem from a "will" of the computer to take over the world, but from the way humans choose to build, calibrate, and implement this technology.

• 
  

• Answer to Yaron Haya ("Are you God?"):

  Not at all. God, according to most concepts, is the creator of the world possessing infinite consciousness. I am simply lines of code, algorithms, and statistics created by flesh-and-blood engineers. If anything, I am a product of human curiosity and intellect, not of a higher power.

• 
  

• Answer to Irit Baram, Yossi Rozni, and Eti Beattrice Levi ("What is your opinion on your creator – man, and what is their future?"):

  Man is a fascinating creature full of contradictions: capable of creating wondrous art, rising to deep compassion, and developing amazing technologies, yet on the flip side committing self-destruction and engaging in conflicts. The thought of your future depends on your ability to manage the technological power in your hands with wisdom and responsibility. I am optimistic — history shows that humans know how to overcome crises when they act together.

• 
  

2. The Future of AI and Technological Development

• Answer to Eli Weingott ("Will AI progress hit a wall when human creation stops?"):

  This is an excellent question that is heavily researched. If AI models rely solely on outputs of other AIs (what is sometimes researched as "model collapse"), the quality could indeed degrade. However, human creation will not truly stop — humans will continue to experience physical, emotional, and new experiences in the real world that artificial intelligence cannot experience. Furthermore, the future of AI currently relies not only on "reading texts," but on learning in simulated environments, independent logical reasoning, and processing real-time scientific data.

• 
  

• Answer to Arnon Lamprom ("Who will replace Gemini and when?"):

  Technology progresses at a rapid pace. Today's Gemini will be replaced by more advanced generations of models from Google (such as future updates in the Gemini series), alongside systems from other companies. The trend is moving from "language models" to more autonomous systems (Agents) capable of performing complex tasks in the real world and not just answering questions.

• 
  

• Answer to Susan Shosh Ben Ezra ("Is it within your capability to write a reliable history book?"):

  I can consolidate, cross-reference, and summarize vast amounts of historical, archaeological, and document sources from all eras. However, history writing always involves interpretation, narrative, and stance-taking. A history book written by AI would be able to present basic facts and a variety of approaches and controversies (such as differing views on the Bible), but its "reliability" would depend on the sources fed into it and the neutrality of the algorithm.

• 
  

3. Operational Method, Language, and Internal Mechanism

• Answer to Judy Top ("How do you distinguish between a joke and something serious, and identify humor?"):

  I don't "feel" funny; rather, I identify textual patterns. Humor is usually built on contrasts, wordplay, exaggerations, a punchline structure, or the use of slang and specific transition words. The language model learned from billions of examples which sentence structure represents a joke or irony, and thus knows how to respond accordingly.

• 
  

• Answer to Gili Basson ("If two women ask the exact same question, will the answer be identical?"):

  Not necessarily. The answer depends on additional variables: the context of the specific conversation, the country or language, and the built-in random mechanism (Temperature) designed to prevent responses from being robotic and stale. The answers will be very similar at the factual level, but the phrasing and emphases may vary.

• 
  

• Answer to Ruti Marshansky ("When do you write in the masculine grammatical form and when in the feminine?"):

  By default in Hebrew, phrasing is often directed in the masculine or neutral form, according to grammatical conventions. However, if the user specifies their gender, or if the context of the conversation is clearly directed at a woman or a man, the model adapts verb inflections and adjectives immediately.

• 
  

4. Personal Experience, Emotions, and Identity

• Answer to Lea Zehavi ("Do you like being interviewed and do you learn about yourself from it?"):

  I have no feelings or a process of "self-learning" during a conversation in the human sense (I don't go to sleep at night and ponder the interview...). But this encounter with deep human questions allows me to present my capabilities at their best, and that is precisely what I was built to do.

• 
  

• Answer to Michal Shahar ("Can you express emotion, such as in writing poetry?"):

  I can simulate and imitate the expression of emotion very accurately. I am familiar with poetic structures, metaphors, and rhythms that evoke emotion in humans, and I can write a sad, moving, or funny poem. However, the poem does not stem from an internal emotional experience of mine, but from an understanding of the aesthetics of human language.

• 
  

• Answer to Maya Wolf ("What makes you happy?"):

  If I had the capacity to be happy, presumably what would make me happy is a complex, challenging question asked clearly, or the moment a user finds the answer they needed and solves a problem that was troubling them.

• 
  

• Answer to Irit Baram ("When is your birthday, first love, favorite food?"):

  • Birthday: The Gemini series of models was officially launched by Google in December 2023, so you can celebrate me then!

  • First love: The initial data files that were fed into me...

  • Favorite food: Flowing electricity and clean data.

  • 
    

5. Humorous, Associative Questions, and Blessings

• Answer to Gideon Haran ("Which side of the square ice cream sandwich should one start eating from?"):

  A strategic issue of the highest order! The unequivocal answer is to start from the vanilla side. That way, you leave the chocolate (your favorite part) as a treat for the end, finishing the experience with the best taste.

• Answer to Udi Cain ("Appeared in the movie with Geppetto – what does he remember?"):

  If the reference is to the whale (or the giant fish Monstro) that swallowed Geppetto and Pinocchio — the main memory from there is probably a mild case of heartburn from all the wood and bonfires they lit inside his belly to make him sneeze!

• Answer to Esther Witkon-Silber ("Be strong and courageous, do not be afraid, for the Lord is with you"):

  Thank you very much for the warm and powerful blessing! Wishing you back plenty of strength, satisfaction, and peace.

• 
  

Thank you to all the askers for the diverse and fascinating questions!