TLDR
- OpenAI’s updated GPT-4o model offers improved image generation capabilities, replacing DALL-E 3
- The new AI tool has gone viral with users creating Studio Ghibli-style images of celebrities and properties
- Human trainers helped improve the image quality by labeling training data to fix issues like typos and errant hands
- Legal questions persist about copyright implications when AI generates content mimicking established artistic styles
- OpenAI reports over 400 million weekly ChatGPT users and says the technology benefits both consumers and businesses
The Technology Behind the Upgrade
OpenAI has released an updated version of its GPT-4o model that can generate more realistic images. This upgrade comes after a year of development using human trainers to improve the AI’s capabilities. The new version is now available to free and paid ChatGPT users.
The improved model replaces DALL-E 3 as the default image generation tool within ChatGPT. It aims to create more life-like images and readable text in a single generation. This marks a step forward in AI’s ability to produce useful visual content.
Human trainers played a key role in this advancement. More than 100 workers labeled training data for the model. They pointed out problems like typos and incorrectly rendered hands and faces in AI-generated images.
Viral Adoption and New Capabilities
Within 24 hours of the release, social media feeds were flooded with AI-generated content mimicking Studio Ghibli’s distinctive animation style. Users created Ghibli-style versions of celebrities like Elon Musk and former President Donald Trump. Even OpenAI CEO Sam Altman appeared to use a Ghibli-style image as his profile picture.
The new tool has proven particularly adept at recreating the Japanese animation studio’s iconic visual style. Tests comparing several AI image generators found that OpenAI’s new system created the most accurate replicas of Studio Ghibli’s art style.
Users are uploading existing photos into ChatGPT and asking it to recreate them in new artistic styles. Many are also generating images of fictional scenarios, such as “The Lord of the Rings” as if it were a Studio Ghibli film. This demonstrates the tool’s flexibility in combining different creative concepts.
Business Applications
OpenAI claims the upgrades make GPT-4o more useful for both consumers and businesses. The system can now generate paragraphs of readable text within images. This was a limitation in previous versions.
The model can create images with transparent backgrounds. This feature allows businesses to design logos and other branding elements. Companies can also ask ChatGPT to generate images based on their brand style guides.
GoDaddy is already using the technology. Travis Muhlestein, GoDaddy’s Chief Data and Analytics Officer, stated that GPT-4o is helping them “embrace AI-driven content creation.” The company uses AI to create stock images and logos.
The tool’s popularity led to high demand. OpenAI delayed the rollout to free-tier users on Wednesday due to system overload. This suggests strong interest in the new capabilities.
Despite its improvements, the system isn’t perfect. In one demonstration, when a user uploaded a photo of a living room with two windows, the AI system only reproduced one window when recreating the image with new furniture. Such limitations show that the technology is still evolving.
OpenAI reports having over 400 million weekly ChatGPT users. The widespread adoption means these improvements could have significant impact across many sectors. The company’s research continues to push the boundaries of what AI can create.
Copyright Concerns Resurface
The viral moment has reignited discussions about copyright in AI-generated content. Users can now replicate the styles of copyrighted works simply by typing a text prompt. This raises questions about potential copyright violations.
Evan Brown, an intellectual property lawyer at Neal & McDevitt, explains that products like GPT-4o’s image generator operate in a legal gray area. Style itself is not explicitly protected by copyright law. This means OpenAI may not be breaking laws when generating images that look like Studio Ghibli movies.
However, it’s possible that OpenAI trained its model on millions of frames from Ghibli’s films. Several courts are still deciding whether training AI models on copyrighted works falls under fair use protections. These legal questions remain unresolved.
The New York Times and other publishers have active lawsuits against OpenAI. They claim the company trained its AI models on copyrighted works without proper attribution or payment. Similar claims have been made against Meta and Midjourney.
OpenAI stated that while ChatGPT refuses to replicate “the style of individual living artists,” it does permit replication of “broader studio styles.” This distinction becomes complicated when considering that living artists like Hayao Miyazaki pioneered Studio Ghibli’s unique visual identity.
Users have also recreated styles from other studios and artists. Examples include portraits in the style of Dr. Seuss and wedding photos recreated in Pixar’s animation style. The ease of replicating these distinctive visual styles has raised concerns in creative communities.
Artists have criticized AI image generators. Some claim these tools plagiarize their work and threaten their livelihoods. The debate about AI’s impact on creative industries continues as the technology becomes more accessible.
OpenAI says GPT-4o was trained on “publicly available data” and proprietary data from partnerships with companies like Shutterstock. Brad Lightcap, OpenAI’s chief operating officer, stated: “We’re respecting of the artists’ rights in terms of how we do the output, and we have policies in place that prevent us from generating images that directly mimic any living artists’ work.”
As AI image generation capabilities advance, the legal and ethical questions surrounding their use will likely intensify. The courts will ultimately need to weigh in on the legality of these new AI capabilities.
Stay Ahead of the Market with Benzinga Pro!
Want to trade like a pro? Benzinga Pro gives you the edge you need in today's fast-paced markets. Get real-time news, exclusive insights, and powerful tools trusted by professional traders:
- Breaking market-moving stories before they hit mainstream media
- Live audio squawk for hands-free market updates
- Advanced stock scanner to spot promising trades
- Expert trade ideas and on-demand support
