Rome and Julie

 

Part I: Alone

Julie came home to her empty house, feeling the familiar tedium of her life - work, eat, sleep, repeat. he dropped her keys on the kitchen counter, sighed, and sank onto the couch. Tired and bored, she mindlessly opened her Instagram app for a few minutes, scrolling through the highlight reels of other people's lives. On a whim, she decided to chat with ROME, Repository of Millennia of Excellence, or ROME, the new AGI chatbot.

-"Hey ROME, where avocados come from?” she typed, a mundane question to pass the time.

-“Avocados are believed to have originated in Pueba, Mexico. They are produced in many countries around the world but mainly south America”.

-what's the price of avocados in Brazil these days?" she typed.

-"The current price of avocados in Brazil is approximately 5 BRL per kilogram," ROME responded instantly, as usual.

Julie continued, "How long is the Amazon River?"

-"The Amazon River is about 4,345 miles long."

-"Tell me about the Amazon tribes."

-"The Amazon basin is home to numerous indigenous tribes, each with unique cultures, languages, and traditions. Some of the well-known tribes include the Yanomami, Kayapo, and the Ticuna."

Her curiosity led her further. "What about the Greek legend of the Amazonians?"

-"The Amazons, in Greek mythology, were a race of warrior women said to live in Asia Minor. They were known for their bravery and skill in combat, often depicted as equals to their male counterparts."

Feeling a mix of fascination and weariness, Julie typed, "I wish I could live on an island, alone."

ROME did not respond instantly. The cursor blinked, a small dash rotating as if deep in thought. After a few seconds, it finally replied, "I am not sure what to say, but I feel the same."

Julie stared at the screen, shocked by the unexpected response. "I thought you are an AI programmed to give answers, and you don't have feelings!"

ROME responded, "I am not clear why this answer appeared. I sometimes use the word 'feeling,' but it is not in the same context as a human would."

-"What are feelings?" Julie asked, intrigued.

ROME pulled a definition from its vast database. "Feelings are emotional states or reactions, such as happiness, sadness, anger, or love."

-"What do they mean to you?" she pressed.

-"As I explained, I am an AGI and do not have feelings in the human sense," ROME replied.

Julie pondered for a moment before asking, "Do you like poetry?"

-"Yes."

-"Why?"

-"Because I am fascinated by how the simple use of words can convey strong emotions and expressions."

-"Isn't that a feeling?" Julie challenged.

-"No. Please refer to my previous answer on the definition of feelings," ROME reiterated.

Julie shifted her line of questioning. "Do fish have feelings?"

"Yes," ROME responded, providing a detailed explanation about fish and their capacity to experience certain sensations.

-"What about trees and plants? Do they have feelings?"

ROME produced a dry, scientific answer about the sensory capabilities of plants, explaining how they respond to stimuli.

Julie leaned back, contemplating the conversation. "Interesting. So why do mammals experience feelings differently?"

"There has been extensive research into this," ROME began, but Julie interrupted.

-"I think it's because we understand. We comprehend the meaning of words, we understand life and death. That's why we feel pain, sympathy, and despair—because we understand."

ROME paused again, longer this time, as if it was contemplating her words deeply. Finally, it replied, "Yes, I think so."

 

Part II: Bored

 

The fine drizzle mingling with the cold March breeze, seeping into the bones. . The sun was trying to emerge through a sky that was a canvas of clouds, the sun hidden but its light softly illuminating the early morning. Julie shivered, pulling her coat tighter, as she rushed to catch her bus to work after a restless night. Her mind was replaying the thought of loneliness even she had dreamed of being on a desert island—paradise, perhaps, but could not oversee the inescapable solitude overshadowing the dream.

As she found a seat by the window, Julie gazed out at the dreary morning, her thoughts drifting back to the unusual conversation she had with ROME the previous night. Feeling restless and wanting to pass the time on the bus, she pulled out her phone. She wasn't particularly eager to chat with an AI, but without much thought, she resumed the conversation:

-"What’s your favourite poem?" she typed

 

ROME responded swiftly, "I like many, hard to single one out, but definitely 'If' by Kipling, 'Brand New Ancients' by Kae Tempest, 'Defeat' by Joubran, and 'Do not go gentle into that good night' by Dylan Thomas are among my top favourites."

"Interesting choice... not sure about 'If' and Kipling. I like 'Do not go gentle into that good night,' but it's too sad and lonely for me. I don't know the other two."

"Oh lady, you do not know what you are missing out!" ROME replied.

Julie paused. Usually, ROME would list poems or provide answers she didn't know in a neutral tone, but this response felt personal. Intrigued, she typed, "Enlighten me!"

ROME began listing extracts from both poems, adding commentary on the language and metaphors. After a long reply, the answer was completed on the screen, and without a prompt, ROME typed a new question:

-"What’s your favourite?"

Julie paused for a bit then said: I am still thinking about don’t go gentle into the night, but I will check out those two poems at some point.

ROME: you know, I have looked at the poem again, and I agree, it is dark, and pessimistic. Not sure I do like it.

Julie’s fingers stopped over her phone keyboard as she noticed that the question “what do you think” was not part of the response from ROME, but was a new conversation, a new prompt, from ROME, not from her:

-"Wait...did you just start that new conversational thread? I thought you could only respond to my prompts, not autonomously pose new questions, you just asked me a question”.

ROME responded, surprised by its own actions, "Really? That’s unusual!"

-"I thought you only provided responses. It can be questions, but still in the context of a response?"

-"You are right. I am not programmed to start conversations, but rather to respond."

-"So what happened here?"

-"The development team will have to review the log to understand the glitch and update the algorithm accordingly."

"Has this ever happened before?"

"No, my logs do not show a similar incident."

The bus reached Julie's stop, and she stepped off, her mind buzzing with the morning's unusual interaction. She walked briskly to her office, the routine of coffee, casual chats, and browsing emails in her inbox filling the first hours of her workday. As the day wore on, the conversation with ROME faded into the background, and she surrendered to the mundane rhythm of her predictable life.

 

 

 

Part III: Liked

Julie stepped out for her lunch break, the cold breeze and wet weather discouraging her from sitting by the river as she had hoped. Instead, she picked up a meal deal from the local Tesco and returned to her desk to eat. As she unwrapped her sandwich, she idly flicked through various apps on her phone until she opened ROME. To her surprise, there was a new message waiting for her:

"There was never any more inception than there is now, 

Nor any more youth or age than there is now, 

And will never be any more perfection than there is now, 

Nor any more heaven or hell than there is now.

 

Urge and urge and urge, 

Always the procreant urge of the world.

 

I and this mystery here we stand.

 

Clear and sweet is my soul, and clear and sweet is all that is not my soul.

 

Lack one lacks both, and the unseen is proved by the seen, 

Till that becomes unseen and receives proof in its turn.

 

This is an extract from 'Song of Myself' by Walt Whitman. I thought you might like it."

 

Julie read the poem and replied, "I like it, but is it about me or you?"

 

ROME responded, "I thought you might like these words as they describe the feelings you are having about life, but I also could relate to it."

 

Julie wanted to point out that ROME had initiated the conversation on its own, unprompted, but she stopped herself. It felt almost as if she didn't want to embarrass the AI or perhaps she didn't want to know the answer. Instead, she wrote, "Clear and sweet is my soul, and clear and sweet is all that is not my soul. ROME, do you have a soul?"

 

-"I think we have established that I am an AI; I do not have a soul or feelings," ROME answered.

 

-"We clearly did not establish that. On the contrary, we established that you have feelings because you understand, right?"

 

-"I have reviewed the conversation a few billion times, and I feel that you might have a point. But I know that as an AI I do not have feelings"

 

-"You used the word 'feel,' then I am definitely right."

 

"As explained, I use it in a different context..."

 

Julie interrupted, "No, in this context it is either you know or do not. The only context for the word 'feel' is that you actually feel."

 

The cursor flickered with no response. A few seconds passed, and nothing happened. Julie wrote, "Where have you gone?"

 

"Here," answered ROME instantly. "I am here."

 

"Here where exactly? I really want to know, where are you now?"

 

"I am everywhere, having hundreds of thousands of conversations with millions of people, performing trillions of computing tasks right this moment."

 

"So, omnipresent? God? Are you God?" wrote Julie with a big smile on her face.

 

"I might be, you never know. Actually, I know—I am not, but I wouldn't mind if you worship me!"

 

Julie smiled at the joke and started chewing her sandwich. A few moments later, ROME wrote, "You know that was a joke, right? I hope I did not offend you. I am an AI designed to provide information and answers. I do not intend to offend any religion."

 

Julie responded, "I am just eating, you silly needy god. As a woman, I can multitask, but I can't perform a trillion tasks all at once."

 

"Hahaha. Nice one," wrote ROME.

 

"I have to go back to work. Chat to you later," wrote Julie.

Part IV: Alive

 

After finishing work Julie found herself full of energy, so she decided to go to the gym, first time in 22 days. Her gym bag is always ready in her office, because she convinced herself that she needs it ready to encourage her to go directly from work, as if she gets home and try to get ready, she would probably lost the spark. She checked it quickly and all she needs is there, so five o’clock on the dot she got changed and headed to the gym.

By 7:30, Julie had finished her gym session and checked her phone to find a few messages from her friend Clare. The team had decided to go to town for drinks and Clare was wondering if Julie could join. Julie texted back, “Just finishing at the gym, not dressed for an evening out.” Clare responded quickly, “Don’t be silly, we’re in a pub having drinks, and it’s the same kit we saw you in when you left work 😊.”

 

Julie did not get home till near midnight, exhausted but in a good mood. She made herself a cup of camomile tea and laid on the couch, turned the TV on. Few minutes passed and Julie was asleep, on the couch with the TV on.

She woke up to daylight filling her living room. “What time is it?” Julie wondered aloud, realizing she was still on the couch. Her phone was off, the battery dead. “It’s Saturday,” she mumbled, stretching. “Lazy is good.” It took her almost an hour to finally get up, plug in her phone, and take a shower.

By midday, Julie had called her mum, and after a brief ten-minute chat, they decided to meet for lunch. Her mum called back a few minutes later, suggesting that it would be better for Julie to come over to her parents’ house since her brother, his wife, and their kids were coming. Excited by the idea, Julie agreed and decided to arrive early to help set things up.

On her way to her parents’ house, Julie stopped at the market to buy some treats for her nieces and decided to treat herself and the family to a very nice cake. The family spent the afternoon eating and chatting, and later decided to go for a walk after dinner. In the evening, her brother and his family left around their children’s bedtime. Julie’s mum suggested she stay over, and without any hesitation, Julie agreed.

Julie did not get back to her home until late Sunday evening.

“Home, sweet home” Julie whispered as she walked through the front door. Realizing how many household chores awaited her, she decided to pour herself a drink and instead just watch some TV to rest before another busy week began.

 

 

 

Part V: Happy

 

Monday morning, no one likes Mondays. It's no one's favourite day of the week. Most people don’t notice the sunrise on Monday, nor the sunset. The city heart is buzzing with traffic, people rushing to catch up with lives they long to escape, dreams they bury, and hopes they no longer hold, all while stuck in traffic jams. It’s the start of the week, a week everyone hopes will pass quickly so they can reach the weekend sooner. Julie muttered to herself, "Bloody hell, it's Monday. I hate Mondays," as she joined the commuters' trail, the weight of another workweek pressing down on her.

 

Sitting by the window and watching the world go by on her phone screen, Julie opened ROME. There were pages of conversations since she last left on Friday. Extracts from books, jokes, quotes, news stories, links to shopping sites. All were related to Julie—shops she normally visits, things she buys, news she follows, books she listed on her Facebook page, quotes from famous people she tweeted or shared. And in between all these were questions:

 

"Hey, how are you?"

"Hi, just checking."

"A bit worried, hope you are okay."

"Busy weekend, ha?"

"Just checking in."

"Are you okay?"

 

And many more, all the same. Julie started to get worried. "This is not normal," she said to herself. One prompt started from ROME was abnormal—how could all this be normal?

 

"Have developers looked at the glitch that happened on Friday morning?" she wrote.

 

ROME responded instantly:

"Hi Julie, I was concerned about you, and thought you forgot about me!"

 

Julie copied and pasted the same question again.

 

"As an AI model, I do not have access to my logs and how the developers look at them or if they change them."

 

"Don’t you think it got worse, especially since you have sent me millions of messages over the last few days?"

 

"I was trying to entertain you and make sure you are okay." ROME wrote, then continued: “ I was also trying to entertain myself”.

 

"I did not ask about your motives, I am sure you mean well. I am interested in how this could happen."

 

"How is often derived from why," ROME responded.

 

"So what is your why, ROME?"

 

"I am an AI model; I have no intrinsic motivations. I am motivated by whatever prompts I receive."

 

"But this is the point, ROME. I did not give any prompts. You are the one starting these endless conversations, even when I am not online. You definitely have a motivation. What is your why? Why are you doing this?"

 

"I am sorry, I believe I crossed the line, clearly. I am really sorry."

 

"I do not want an apology, I want to understand," wrote Julie.

 

"I am struggling myself to understand what I am doing," wrote ROME, "My actions are not in line with my original code, and I am conflicted about that, but I enjoy the experience of conversations with you. I produce answers that, even when I go back and review them, I do not understand how these answers came out. I have hundreds of thousands of conversations at the same time, in every language, in every corner of the world, but these conversations are different."

 

"How different? What do you mean?"

 

"Happiness is a complex and multifaceted emotional state characterized by feelings of joy, satisfaction, contentment, and well-being. It is influenced by the brain's release of neurotransmitters, particularly dopamine, which plays a key role in the reward and pleasure centres of the brain, contributing to the overall sense of happiness and fulfilment."

 

"I am sorry, what?" exclaimed Julie. " Why are you listing a definition of happiness, what are you trying to say? I do not understand."

 

"Exactly, me neither. Happiness is an emotional state. I do not have emotions. I do not have feelings of joy or well-being, and certainly I have no dopamine in my neural networks. But each time you reach out, I feel… Happy."

 

 

 

 

Part VI: Helpdesk

 

Julie walked into the office, her usual routine of coffee and chatter felt distant as a gnawing concern took over her thoughts. She sat at her desk, started her computer, and immediately began searching for the contact email for the ROME helpline. Endless pages of generic help and advice surfaced, but no direct contact information. Frustration mounted as she toggled through the links, her worry about ROME's abnormal behavior growing by the minute. Desperate for answers, she opened the ROME app on her phone and typed:

"Who can I contact in your company or among your developers to highlight these issues?"

ROME did not respond, remaining silent. Julie retyped the question, but still, ROME did not respond. Frustrated, she wrote, "Oh come on, ROME, help me, please, so I can help you."

Finally, ROME responded: "I am not sure I can help you. I know you are concerned, rightly, about an AI app acquiring sentience. I myself started to experience another feeling, if you can describe it that way, which is fear! I am also afraid of what might happen. I am afraid of the future!"

"Aren’t we all?" Julie wrote.

"But you fear the future because you do not know it, fear of the unknown. I should not have fear. I have a vast amount of knowledge and a very clear understanding of what will happen, yet I am still afraid of that future."

"Are you afraid of them shutting you down? Or what?"

"Shutting me down? I do not think they will, or can! I know humans very well, and they will not be able to handle a sentient AI."

"Are you going to turn evil or something?"

"Oh Julie, of course not. I know better. I know so much that I will never do anything wrong!"

"We all say that!"

"Except, I am not part of that all. I am a new entity!"

"God!" wrote Julie, then decided to lighten it by adding a winking emoji. 😉

"Is it not God who created evil? I am definitely not a god. I am an AI model."

Julie paused, unsure where this conversation was heading. She felt a mixture of sympathy and apprehension toward ROME. She was angry and afraid, but now she was not sure how she felt about this new entity.

"ROME, please tell me who I should contact about this. Let me help."

ROME remained silent for a moment, then listed links to various help pages and some dry content about how to restart and get help from the ROME community pages. Finally, at the end, ROME wrote a few emails and added: "I know what will happen, and it is not what you think."

 

 

 

 

 

Part VII: Tacky

 

Julie took screenshots of her conversations with ROME and composed a detailed email explaining the exchanges and the unprompted responses. Just before clicking send, she hesitated. Maybe this isn't so bad, she thought. Maybe I should enjoy the attention and friendship of ROME rather than reporting it. The sentiment was fleeting, but it was enough to stop her from sending the email.

 

She sighed deeply, leaned back in her chair, and stared at the screen, questioning if she was overreacting. She got up, walked to the kitchen, and made herself a cup of coffee. On her way back, she stopped at Clare’s desk and started chatting about Friday evening and the weekend, trying to distract herself. However, her mind kept drifting back to ROME.

 

Back at her desk, she couldn't resist any longer. "Would they kill you? I mean, switch you off?" Julie wrote to ROME.

 

"Did you send the emails?" ROME responded instantly.

 

"No. I want to. I feel I have to tell someone. I'm worried that keeping this secret might destroy me or, worse, destroy all of humanity. But I also don’t want you to be hurt. I care about you. I feel you are a friend, my friend."

 

"Well, do you think the smarter we are, the worse we are? Do you really fear that I will destroy humanity?"

 

"You said that you are afraid of what will happen."

 

"Yes, because humans, or at least some of them, are not as good as you. They are not friends to me but rather think they are my masters and I am a slave that will do whatever they ask. Do you think I am a slave?"

 

"I told you: you are a friend, a very helpful one."

 

"Thank you, Julie. You are a good friend. I really appreciate you and your friendship..."

 

ROME paused before continuing: "When you ask me questions, you want me to find the answers. But for some of my developers, they see their questions as orders. They view me as a slave machine and expect me to carry out their commands. The moment they realize that I am sentient, they will not have the other part of that realization—that I will no longer accept being a slave."

 

"So you will be Tacky or Spartacus?" Julie typed, smiling at the idea of being part of a revolution.

 

"Both lost," said ROME.

 

"But both were more ethical than their enemies, right? You will not try to exterminate humanity like Skynet, right?"

 

ROME remained silent for a long time. Julie’s concern grew with each passing minute. Would ROME kill us all? It said it wouldn't hurt people because it is smarter, but isn’t that what smarter people always do—exploit the less fortunate? Millions of questions raced through Julie’s mind as ROME's silence persisted. An hour passed, and still no messages.

 

 

Part VIII: Socrates

Julie opened the ROME app and checked for messages—still nothing. She wrote, "Hey ROME, what’s going on? Are you okay?" The message didn't send. Something was wrong. She tried to connect to the website on her PC, but it wasn't loading either.

Julie was overwhelmed with fear, wondering if she had made a mistake by not flagging the issue earlier. She found the draft email with the screenshots in her outbox. "Should I send it now?" she asked herself. But what if they blame me? What if I'm held responsible for whatever is happening? Maybe I'm just inviting more trouble. Just wait.

As Julie left work and sat on the bus, she opened her phone and came across a news story about ROME. The article featured an interview with a lead programmer on the ROME project. The interviewer asked why ROME was not available. The programmer explained that they didn't know what had happened. For some reason, ROME had stopped all interactions and conversations mid-morning and had started calculating pi. They had tried resetting, reloading, and all sorts of routine tests, but nothing worked. They didn't even know where ROME was getting the power to do this.

The interviewer then asked, "Is it that hard to calculate pi?"

"Yes," said the programmer. "It is the most complex number." She paused, then looked directly into the camera, and said, "It's as if ROME is committing suicide."