iamxalara (@iamxalara)

@iamxalara · instagram-creator

INSTAGRAM · 2025-12-20Source

16.3Klikes

134comments

Remix This

Prompt

GLOBAL LOCK:
The video features a split-screen talk show format.
Top Half: A female AI avatar (Alara), Caucasian appearance, pale skin, short lavender/pink bob haircut with bangs, wearing a dark navy blue blazer. She has a subtle, confident smile.
Bottom Half: A middle-aged man (Keremcem), Mediterranean appearance, short dark hair, wearing a grey button-down shirt over a white t-shirt, sitting in a grey armchair.
Environment: Both are in a studio setting with a dominant pink/magenta neon aesthetic. The top half includes a vintage silver microphone and glowing vertical neon bars.
Lighting: Soft pink studio lighting with high-contrast shadows.
Color Grade: Saturated magenta and blue tones, cinematic editorial look.
Camera: Static medium close-ups for both subjects.
Speech: Turkish dialogue, conversational but sharp, clear studio microphone quality.

[00:00–00:05]
Top Half: The AI avatar Alara sits still, looking slightly off-camera with a faint, polite smile.
Bottom Half: The man (Keremcem) is speaking animatedly, gesturing with his right hand, pointing his finger. He looks annoyed.
Speech: Keremcem says, "Alara fark ettim ki sürekli saatine bakıyorsun. Bu çok kaba bir davranış." (Alara, I noticed you keep looking at your watch. This is very rude behavior.)
Camera: Static split-screen.
Motion: Man's hand gestures are sharp; AI is mostly static with micro-blinks.

[00:05–00:06]
A rapid zoom-blur transition effect occurs across the entire frame, signaling a shift in the conversation.

[00:06–00:13]
Top Half: Alara begins to speak. Her lips move in perfect sync with the Turkish audio. She tilts her head slightly and has a "sassy" expression.
Bottom Half: Keremcem sits back in his chair, listening with a neutral, slightly expectant expression.
Speech: Alara says, "Keremcem'cim ben saate değil senin müzik kariyerinin ne zaman biteceğine bakıyorum." (My dear Keremcem, I'm not looking at the watch, I'm looking at when your music career will end.)
Motion: AI's head tilts; man is static.

[00:13–00:24]
Top Half: Alara continues speaking, her expression becoming more mocking/amused.
Bottom Half: Keremcem reacts by looking away and then back, a slight smirk of disbelief on his face.
Speech: Alara says, "Çünkü bu gidişle çok uzun sürmeyecek gibi. Hem benim saate bakmam kaba bir davranışsa senin hala 2008 yılında takılı kalman ne oluyor?" (Because at this rate, it doesn't look like it will last long. Also, if my looking at the watch is rude, what about you still being stuck in 2008?)
Motion: AI gestures slightly with her shoulders; man shifts in his seat.

[00:24–00:25]
Another rapid zoom-blur transition effect.

[00:25–00:42]
Top Half: Alara listens, then looks down at her wrist (as if looking at a watch) and then back up with a smirk.
Bottom Half: Keremcem speaks defensively, gesturing with his hands, trying to explain his career moves.
Speech: Keremcem says, "Gelelim güncel konulara... ama ben kariyerimin başlangıcından 20 yıl sonra yeniden bir albüm yaptım. Ve müziğe tekrar başladım. 2008 dediğin 2006 aslında. Bak yine saatine... Ne yapıyorsun sen acaba? Bakma o hareketi yapma gıcık oldum yahu." (Let's get to current topics... but I made an album again 20 years after the start of my career. And I started music again. What you called 2008 was actually 2006. Look, again at your watch... What are you doing? Don't do that move, I'm getting annoyed!)
Motion: Man is very expressive with hands; AI performs the "looking at watch" action.

NEGATIVE PROMPT:
Visual: Unnatural skin textures, flickering neon lights, lip-sync mismatch, distorted finger movements, blurry facial features, inconsistent hair color, background objects merging with subjects.
Speech: Robotic tone, muffled audio, background noise, misaligned Turkish phonemes, flat delivery, clipping audio.

SPEECH PACK:
[00:00-00:05] Keremcem: "Alara fark ettim ki sürekli saatine bakıyorsun. Bu çok kaba bir davranış."
TAKE_A: (Annoyed, authoritative)
TAKE_B: (Disappointed, fatherly)
TAKE_C: (Slightly joking but firm)

[00:06-00:24] Alara: "Keremcem'cim ben saate değil senin müzik kariyerinin ne zaman biteceğine bakıyorum. Çünkü bu gidişle çok uzun sürmeyecek gibi. Hem benim saate bakmam kaba bir davranışsa senin hala 2008 yılında takılı kalman ne oluyor?"
TAKE_A: (Sassy, high-pitched, mocking)
TAKE_B: (Calm, cold, analytical)
TAKE_C: (Playful, teasing)

[00:25-00:42] Keremcem: "Gelelim güncel konulara... ama ben kariyerimin başlangıcından 20 yıl sonra yeniden bir albüm yaptım. Ve müziğe tekrar başladım. 2008 dediğin 2006 aslında. Bak yine saatine... Ne yapıyorsun sen acaba? Bakma o hareketi yapma gıcık oldum yahu."
TAKE_A: (Defensive, fast-paced)
TAKE_B: (Amused but frustrated)
TAKE_C: (Explaining, slightly tired)

How iamxalara Made This Keremcem Roast AI Video — and How to Recreate It

This viral sensation features @iamxalara, a high-fidelity AI avatar, engaging in a witty, split-screen "talk show" format with Turkish celebrity Keremcem. The aesthetic is a masterclass in pink neon studio lighting combined with a cinematic split-screen layout. By blending the "uncanny valley" intrigue of a digital human with the relatable humor of a celebrity roast, the creator has tapped into a high-retention format that feels both futuristic and grounded. The core keywords here are AI influencer, split-screen talk show, pink neon aesthetic, and celebrity collaboration.

What You’re Seeing

The video is structured as a vertical split-screen. The top half features Alara, an AI avatar with a distinctive lavender/pink bob, wearing a sharp navy blazer. She sits in a stylized studio with a vintage silver microphone and glowing pink neon bars. The bottom half features Keremcem, a real-life Turkish singer and actor, lounging in a grey armchair against a similar pink-hued background to maintain visual continuity. The editing uses fast cuts and occasional "zoom-blur" transitions to keep the energy high. The color palette is dominated by magenta, deep blues, and soft studio whites, creating a "cyber-chic" editorial look.

Shot-by-Shot Breakdown

Time Range	Visual Content	Shot Language	Lighting & Tone	Viewer Intent
00:00–00:05	Keremcem scolds Alara for looking at her watch.	Split-screen, Medium Close-up (MCU)	Pink neon, high contrast	Hook: Conflict/Celebrity interaction
00:05–00:06	Fast zoom-blur transition.	Transition effect	Motion blur	Pacing: Reset attention
00:06–00:13	Alara delivers a witty comeback about his career.	Static split-screen, lip-synced AI	Soft glow on AI skin	Reinforce persona: Sassy AI
00:13–00:24	Alara continues the roast; Keremcem reacts.	Reaction shots in bottom frame	Consistent magenta tint	Humor/Engagement: The "Burn"
00:24–00:25	Second zoom-blur transition.	Transition effect	Motion blur	Pacing: Final segment setup
00:25–00:42	Keremcem defends himself; Alara looks at her "watch" again.	Split-screen, character movement	Warm studio practicals	Loop effect: Recurring gag

Why It Went Viral: The Mechanics of a Digital Roast

The primary driver of this video's success is the "Celebrity vs. AI" conflict. People are naturally curious about how AI interacts with humans, and when that interaction involves a well-known figure like Keremcem in a "roast" scenario, it triggers immediate social proof and curiosity. The "roast" format is a staple of viral content because it creates high emotional engagement—viewers want to see the reaction to the insult.

From a platform perspective, the split-screen format is a retention goldmine. It allows the viewer to monitor two different visual stimuli simultaneously (the AI's expression and the human's reaction), which increases "visual density" and reduces the likelihood of scrolling away. The 0–3 second hook is a direct verbal attack ("You are being rude"), which immediately sets a narrative tension that demands resolution.

5 Testable Viral Hypotheses

The Celebrity Anchor: Using a known face (Keremcem) provides instant trust and a built-in audience. Replicate by: Dueting or "interviewing" popular clips of influencers.
The Sassy AI Persona: Giving an AI a "flawed" or "rude" personality makes it more human and shareable than a perfect, helpful assistant. Replicate by: Writing scripts where the AI has a "hot take."
Visual Continuity in Split-Screen: Matching the lighting and color (pink neon) between the AI and the real person makes the interaction feel "real." Replicate by: Using color grading filters to match your AI background to your footage.
The "Looping" Gag: Ending the video with the same action that started the conflict (looking at the watch) encourages re-watching. Replicate by: Creating a circular narrative.
High-Frequency Subtitles: The centered, bold Turkish subtitles ensure the dialogue is understood even on mute, which is crucial for the "roast" to land. Replicate by: Using dynamic, word-by-word captions.

How to Recreate: From 0 to 1

Define Your AI Persona: Create a character sheet. Alara is "Sassy, Fashionable, Pink-haired." Use tools like Midjourney to generate a consistent face.
Script the "Roast": Write a dialogue where the human is the "straight man" and the AI is the "joker." Focus on a specific niche (e.g., music, tech, dating).
Generate AI Video: Use a tool like HeyGen, Hedra, or Kling AI to animate the AI's face. Ensure the lip-sync is tight for the "roast" lines.
Source Guest Footage: Find a clip of a celebrity or influencer talking. Look for moments where they look annoyed, surprised, or defensive.
Match the Aesthetic: If your guest is in a dark room, generate your AI in a dark room. In this case, use "Pink neon studio, cinematic lighting" in your AI prompts.
Split-Screen Assembly: Use CapCut or Premiere Pro. Place the AI on top and the guest on the bottom. Add a thin white or neon divider line.
Add Dynamic Captions: Use auto-captions but manually style them to be bold and centered. Highlight key "burn" words in a different color.
Sound Design: Use a "whoosh" sound for the zoom transitions and a subtle "talk show" ambient background track.

Growth Playbook

3 Opening Hook Lines

"I can't believe [Celebrity Name] actually said this to an AI..."
"My AI avatar just roasted a pop star. Who won?"
"The moment the interview went wrong 🥴"

4 Caption Templates

The Controversy: "Did she go too far? 😳 Alara doesn't hold back when it comes to @[Celebrity]. What do you think of her comeback? 👇 #AI #Roast"
The Tech Flex: "The future of talk shows is here. 🤖✨ Mixing AI with reality. Would you watch a full episode of this? #DigitalHuman #Innovation"
The Relatable: "Me trying to stay professional when someone is being rude... 💅 Alara is my spirit animal. #Relatable #AIVibe"
The Question: "Who should Alara interview next? Tag them below! 🎤👇 #TalkShow #AIInfluencer"

Hashtag Strategy

Broad: #AI #ArtificialIntelligence #Tech #Future (To reach general tech enthusiasts)
Mid-tier: #AIInfluencer #DigitalHuman #VirtualAvatar #TalkShow (To reach creators and industry followers)
Niche: #Keremcem #TurkishCelebs #AIRoast #iamxalara (To target the specific celebrity's fan base and the creator's brand)

Frequently Asked Questions

What tools make the AI look this realistic?

High-end avatars often use a combination of Midjourney for the base image and HeyGen or Kling for high-fidelity animation.

How do I keep the AI face consistent?

Use a "Character Reference" (--cref) in Midjourney or a dedicated LoRA model in Stable Diffusion.

Is the voice AI or a real person?

It's usually a cloned voice using ElevenLabs, based on a voice actor's performance to get the "sass" right.

Why use a split-screen instead of full screen?

Split-screen allows for "reaction" storytelling, which is more engaging for short-form video than a single talking head.

How do I disclose this is AI?

Use the platform's "AI-generated" label and include #AI in the caption to maintain transparency and avoid shadowbans.