**Explosive Growth in Image Generation**
The trend of using OpenAI’s AI chatbot ‘ChatGPT’ to transform real-life photos into cartoon styles is reshaping the AI industry. OpenAI sparked a wave of generative AI enthusiasm by launching ChatGPT at the end of 2022. Following this, big tech companies like Google and Meta engaged in a competitive race to improve AI capabilities, investing astronomical sums. However, AI was rarely utilized in daily life. The recent trend of transforming photos into animation styles like Studio Ghibli or Disney using ChatGPT’s image generation capability has prompted widespread use of AI, marking its entrance into the mainstream. Paid subscriptions for this feature have also increased, paving the way for AI monetization.
**Surpassing 20 Million Paid Subscribers**
On March 27th, OpenAI integrated an image generation function into ‘ChatGPT-4o’. This feature modifies or creates new images based on textual commands. On the day it launched, Sam Altman, CEO of OpenAI, shared an image of himself altered to resemble a Ghibli animation character on X (formerly Twitter). Garnering praise for being ‘cute’ and ‘warm’, users began altering their photos to animation styles like ‘Ghibli’, ‘Disney’, and ‘Simpsons’, and sharing them on social media. This resulted in a surge of ChatGPT users. According to external sources like The Information, as of the end of last month, ChatGPT had over 500 million weekly active users (WAU) and surpassed 20 million paid subscribers. Altman mentioned, “It used to take five days to increase users by a million when we launched ChatGPT 26 months ago, but now it happens in just one hour.”
Market research firm Sensor Tower reported that following the addition of the advanced image generation feature, ChatGPT’s global app downloads and weekly active users increased by 11% and 5%, respectively, from the previous week.
With GPT-4o, free users can generate up to three images per day. However, to use the image generation feature more freely, many have subscribed to the paid version of ChatGPT. The number of paid users increased by 6% within a week after the feature’s launch. Considering it took about 20 months to secure 15.5 million paying customers, this growth rate is significant. An industry insider noted, “There has been a major shift, as before, only those with work-related needs would pay for high-end models.”
With OpenAI leading AI’s popularization and monetization through image generation, a competitive race in multimodal AI is anticipated. Multimodal AI can understand and process diverse data formats beyond language, including images, speech, and videos. Google has enhanced its multimodal function with the release of the ‘Gemini 2.0’ model last December, continuing updates this year. xAI, owned by Elon Musk, acquired the AI video creation technology startup ‘HotShot’ last month, likely aiming to strengthen its image and video functionalities.
**Challenges with Infrastructure and Copyright**
While the ‘Ghibli-style’ trend has paved the way for AI popularization and monetization, it also presents new challenges for AI companies. Compared to text, images have larger data sizes, prompting companies to invest more in AI infrastructure like data centers. It’s reported that graphic processing units (GPUs) required for image and video generation use up to ten times more than those for text-based information retrieval. An insider indicated, “Given that the price for ChatGPT’s paid versions, Plus and Pro, is 20 dollars and 200 dollars respectively, it may be hard to cover GPU costs.”
Issues related to copyright have also surfaced as ChatGPT mimics Ghibli or Disney artistic styles. The applicability of copyright law to AI-generated Ghibli-style images is ambiguous. In the U.S. and Japan, copyright laws typically protect specific expressions rather than general styles. However, if OpenAI used copyrighted works for training without authorization, it could pose a problem.
**🌟 ChatGPT Ghibli Style**
By the end of last month, when OpenAI added the image generation feature to ChatGPT-4o, a craze began with users converting their photos into the style of Japan’s animation studio ‘Studio Ghibli’. Known for works like ‘My Neighbor Totoro’ and ‘Spirited Away’, Ghibli is renowned for its soft colors and lines that depict characters cutely and gently.