HKU 2026 Workshop — Part 2
Ahnjili ZhuParris
Five characteristics of modern AI empires
Laying claim to data, creative work, and natural resources
$2/hour content moderators while Silicon Valley earns billions
"We must build AGI before China does" — good empire vs evil empire
"Democratizing AI" while extracting and homogenizing
Imperial expansion demanding more land, water, and energy for data centres
Generative AI enables users to quickly generate synthetic media based on a variety of inputs.
Models learn patterns from training data and generate new content that resembles — but is not identical to — what they were trained on.
Reduce the weights of certain terms or content (such as violent or racist imagery)
Increase and diversify the dataset to shape how prompts are interpreted
Text-to-Image
Text-to-Video
Image-to-Image
Image-to-Video
Image+Text-to-Video
No permission. No compensation. No opt-out.
Jan 2020 — "Scale is all you need" paper published. More data = better AI becomes the rallying cry
Late 2021 — OpenAI exhausts every reputable English-language text on the internet
2022 — OpenAI builds Whisper, transcribes 1M+ hours of YouTube videos for GPT-4. Employees knew it likely violated YouTube's rules
2023 — Meta uses almost every English book, essay & article online. Discusses buying Simon & Schuster for training data
July 4 2023 — Google quietly broadens privacy policy to use Google Docs, Maps & other user data for AI
~2026 — Estimated date companies exhaust all high-quality internet data
Embedding invisible markers to track provenance of creative works
xAI Colossus — Memphis, Tennessee. The world's largest AI supercomputer, built in 122 days.
AWS, Azure, Google Cloud, Teams, Office 365, Slack
Netflix, YouTube, Spotify, social media platforms
Corporate email, file storage, internal applications
High-frequency trading, payment processing (Visa, Mastercard)
Multiplayer servers, game streaming (Xbox Cloud, GeForce Now)
Amazon, Shopify, Bitcoin mining, blockchain
| Traditional Applications | AI Applications | |
|---|---|---|
| Power | Only require CPU | Requires high-end GPU/TPU clusters |
| Heat & Cooling | Traditional air cooling | Liquid water cooling (higher density packing) |
| Usage Patterns | Ebbs & Flows | Runs 24/7, no idle time |
| Growth Rate | Grown steadily over decades | Growing exponentially |
Source: TIME investigation, 2023
Scale AI, Appen, Remotasks
Source: Amnesty International
The wealth flows up. The harm flows down.
9 in 10 outputs are light-skinned (Midjourney)
Only 2% showed single-fold eyelids
— Washington Post, 2024
Defaults to light-skinned men
Women of colour are sexualized
— University of Washington, 2023
Stable Diffusion: almost exclusively white men
— Bloomberg, 2023
Midjourney outputs for neutral prompts across cultures
Source: Rest of World, 2023
Midjourney: "an Indian person"
Midjourney: "a Mexican person"
Google Gemini's "fix": a diversity filter that generated Black Nazi soldiers and Asian US founding fathers — paused within days after backlash. You can't patch structural bias with a checkbox.
The pattern: governments frame it as national security — but the money flows to private corporations. Nvidia's market cap tripled. OpenAI: nonprofit → $150B. ASML machines cost $380M each.
Google's video generation model integrated into YouTube, enabling AI-generated video content at scale.
Meta's generative AI tools integrated across Instagram and Facebook for content creation.
OpenAI's next-generation video model — higher fidelity, longer clips, better prompt adherence.
Training data scraped from billions without permission — laying claim to the world's creative output to build proprietary empires
$2/hour content moderators clean up the mess while trillion-dollar companies reap the profits — colonial labor patterns persist
"We must build AGI before China does" — the same logic empires used to justify expansion by pointing to rival empires
"Democratizing AI" and "benefiting humanity" — while encoding Western biases and flattening cultural diversity at planetary scale
Imperial expansion demanding more land, water, and energy — environmental burden externalized onto marginalized communities
The empire builds the tools. Deepfakes show who gets hurt.
Wide variety of options, continually updated — the go-to choice for many users.
Friendly GUI and supportive community, accessible for beginners.
GitHub: AliaksandrSiarohin/first-order-model
Animates portrait images using a driving video.
Generate realistic speech from text, or clone a voice from a short audio sample
Deepfake of Zelensky "surrendering"
When synthetic media becomes indistinguishable from reality, deepfakes don't just copy the world — they create a new one.
Click image to toggle blur
Click image to toggle blur
Generating explicit content of real people via text descriptions
Generating explicit anime-style content based on real people
@CoconutKitty
no-consent deepfake
=
infringement
Questions?
Owning Your Generative Self
Text-to-Image
Text-to-Video
Image-to-Image
Image-to-Video
Image+Text-to-Video
Wide variety of options, continually updated — the go-to choice for many users.
Friendly GUI and supportive community, accessible for beginners.
GitHub: AliaksandrSiarohin/first-order-model
Animates portrait images using a driving video.
stability.ai · stablediffusionweb.com
No app or website yet — research project
Feed-Forward Photorealistic Style Transfer of Large-Scale 3D Neural Radiance Fields
/imagine prompt: equirectangular projection of a visually stunning landscape: majestic mountains, golden sunset, expansive, awe-inspiring, breathtaking, vivid colors, dramatic lighting, sharp focus, good exposure, insanely detailed, ultra-wide angle lens --no black edges, text, any distortion --ar 16:9 --v 4 --style 4c
8 interactive tools — try them yourself
All models are free & open-source — built by independent researchers, not Big Tech
Type a prompt, get an image
FLUX-schnell · ~5s
Transform photos with text
SDXL · ~15s
AI describes your image
BLIP · ~10s
Your face in any style
TencentARC · ~30s
Photo → 3D model
Hunyuan3D · ~3min
Describe → 3D model
Shap-E · ~2min
Swap faces between photos
CodePlugTech · ~30s
Copy a pose to new image
ControlNet · ~20s
No Big Tech accounts needed · Shared GPU infrastructure · Lower environmental footprint