The most recent iteration of the wildly successful artificial intelligence chatbot ChatGPT, GPT-4, was just launched by OpenAI.
The new model can respond to images, for example, by producing tags and descriptions and suggesting recipes based on images of the ingredients.
It can process 25,000 words in total, which is around eight times as much as ChatGPT.
- Advertisement -
Since its debut in November 2022, ChatGPT has been used by millions of users.
Although schools advise against using it, it is frequently requested for producing songs, poems, marketing copy, computer code, and assistance with schoolwork.
ChatGPT uses the internet as it was in 2021 as its knowledge base and responds to queries in language that is as natural and human-like as possible. It may also imitate different writing styles, such as those of authors and songwriters.
There are worries that technology might one day replace a lot of the tasks currently carried out by people.
OpenAI claimed to have worked on GPT-4’s safety features for six months and trained it using user feedback. It did, however, issue a warning that it might still be prone to disinformation dissemination.
- Advertisement -
Initial access to GPT-4 will be granted to ChatGPT Plus members, who pay $20 monthly for a tier of service.
Microsoft’s Bing search engine platform is already powered by it. The tech titan gave OpenAI a $10 billion investment.
In a live demo, it produced a response to a challenging tax question, but there was no way to validate it.
GPT-4 is a form of generative artificial intelligence, much like ChatGPT. Predictive text and algorithms are used by generative AI to produce fresh content in response to commands.
GPT-4 explained
GPT-4 is more innovative and team-oriented than ever. While working with users on creative and technical writing activities like songwriting, screenwriting, or figuring out a user’s writing style, it can generate, edit, and iterate with them.
By scoring in higher approximate percentiles among test-takers, GPT-4 surpasses ChatGPT.
Safety & alignment
Training with human feedback
Open Ai incorporated more human feedback, including feedback submitted by ChatGPT users, to improve GPT-4’s behavior. Open AI also worked with over 50 experts for early feedback in domains including AI safety and security.
Continuous improvement from real-world use
Open AI has applied lessons from real-world use of our previous models into GPT-4’s safety research and monitoring system. Like ChatGPT, Openai will be updating and improving GPT-4 at a regular cadence as more people use it.
GPT-4-assisted safety research
GPT-4’s advanced reasoning and instruction-following capabilities expedited our safety work. Open Ai used GPT-4 to help create training data for model fine-tuning and iterate on classifiers across training, evaluations, and monitoring.
Research
GPT-4 is the latest milestone in OpenAI’s effort in scaling up deep learning.
Infrastructure
GPT-4 was trained on Microsoft Azure AI supercomputers. Azure’s AI-optimized infrastructure also allows us to deliver GPT-4 to users around the world.
Limitations
GPT-4 still has many known limitations that OpenAI is working to address, such as social biases, hallucinations, and adversarial prompts. We encourage and facilitate transparency, user education, and wider AI literacy as society adopts these models. OpenAI also aim to expand the avenues of input people have in shaping our models.
Availability
GPT-4 is available on ChatGPT Plus and as an API for developers to build applications and services.
