Tech Desk 
Feb 16, 2025
In the wake of the artificial intelligence (AI) battle intensifying and evolving rapidly, ByteDance, the renowned Chinese tech giant that owns the short-form video platform TikTok, has launched its Goku AI, an enormously capable AI tool for image and video generation.
Named after the much-famed anime character Goku, Goku AI is said to be rolled out by the TikTok parent in a bid to challenge OpenAI’s dominance in the realm of AI. Creating high-quality, lifelike visuals from mere text prompts, it is said to be an arch-rival to OpenAI’s Sora.
The tool comes on the heels of AI's never-ending evolution and concerns regarding its regulations, accessibility and impact on the world.
How Goku AI works?
The new Chinese AI tool functions utilising rectified flow transformers, an AI model designed for meticulously improving the flow of information when tasked with image or video generation. The component helps Goku AI generate smoother and more accurate visuals.
What sets it apart from typical AI models is that it fine-tunes its digital "brushstrokes" to keep generated visuals free from glitches and distortion.
“Having the best AI model is no longer enough. Goku reinforces that regulation alone won’t determine success—adaptability and application will,” said Lutz Finger, an AI expert and Cornell faculty member, according to Reuters.
Other notable aspects of Goku AI giving it a winning edge over other image and video generators revolve around its open-source nature, following in the footsteps of models like Meta’s Llama and DeepSeek.
