OpenAI – ChatGPT API , a San Francisco-based AI start-up, has recently launched an API for ChatGPT, a text-generating AI with an estimated over 100 million monthly active users. OpenAI also introduced dedicated capacity for enterprise customers.
ChatGPT has gained popularity and has been used to write hundreds of ebooks, coauthor at least one scientific paper, and attracted major media attention. To monetize ChatGPT, OpenAI introduced a premium service, ChatGPT Plus, in February.
Now, with the ChatGPT API, any business can incorporate the AI model into their apps, websites, products, and services.
The ChatGPT API is powered by GPT-3.5-turbo, the most potent text-generating model offered by OpenAI. The API costs $0.0002 per 1,000 tokens or about 750 words, with early adopters including Instacart, Quizlet, Shopify, and Snap.
While the API was initially intended to cut down ChatGPT’s compute costs, it has been improved to make it more accessible and usable. The API underpins Snap’s recently announced My AI chatbot, Quizlet’s Q-Chat virtual tutor, Shopify’s personalized assistant for shopping recommendations, and Instacart’s Ask Instacart tool.
ChatGPT was initially trained on biased data and has been susceptible to prompt-based attacks, but OpenAI has been addressing these issues to mitigate them. The ChatGPT API is now available for any business to leverage ChatGPT’s capabilities in building their AI-powered apps, products, and services.
Highlights
- Monetizing ChatGPT through ChatGPT Plus and the ChatGPT API
- GPT-3.5-turbo: the powerhouse behind the ChatGPT API
- ChatGPT API Early Adopters: Instacart, Quizlet, Shopify, and Snap
- ChatGPT’s Susceptibility to Attacks and Bias
- Making ChatGPT’s Capabilities Accessible to Any Business
In addition to these measures, OpenAI has also restricted access to ChatGPT’s full capabilities to enterprise customers who go through a rigorous vetting process.
These customers receive dedicated compute capacity and access to specialized support to ensure they’re using ChatGPT in a responsible manner.
As for the API, OpenAI says it’s working hard to address the potential for misuse. For example, the company will monitor usage patterns and has implemented a feedback loop so that the AI model can learn from errors and avoid making similar mistakes in the future. Additionally, OpenAI has provided guidelines to developers on how to use the API responsibly and avoid harmful applications.
Despite these challenges, OpenAI remains committed to making ChatGPT available to businesses of all sizes. The company sees enormous potential in the technology, from powering chatbots and virtual assistants to generating automated content and even helping with scientific research.
As Brockman put it, “We want to enable people to build powerful AI systems without needing to be an AI expert.” With the launch of the ChatGPT API, OpenAI has taken a significant step toward achieving that goal.
Whether it will succeed in making AI accessible to the masses remains to be seen, but there’s no doubt that ChatGPT has already made a significant impact on the field of natural language processing and is poised to transform the way we interact with technology in the years to come.