OpenAI, a startup supported by Microsoft Corp (MSFT.O), announced on Tuesday the commencement of the release process for its advanced artificial intelligence model, GPT-4, Reuters reports.
This development marks the beginning of a new era for human-like AI technology and intensifies the competition between Microsoft and Alphabet Inc’s (GOOGL.O) Google.
The company behind the chatbot phenomenon, ChatGPT, revealed in a blog post that its latest innovation is “multimodal,” allowing both images and text prompts to trigger content generation.
The text-input feature will be accessible to ChatGPT Plus subscribers and software developers via a waitlist, while the image-input capability remains in the research preview phase.
This highly anticipated release highlights the increasing reliance of office workers on AI for various tasks and demonstrates the ongoing competitive efforts of tech giants to capitalize on such advancements.
Google, a subsidiary of Alphabet Inc (GOOGL.O), unveiled a “magic wand” for its collaborative software, capable of drafting virtually any document, just days before Microsoft’s expected announcement of AI integration in its Word processor, possibly utilizing OpenAI technology.
A Microsoft executive also mentioned that GPT-4 is now contributing to the power of its Bing search engine.
According to OpenAI, its new technology offers a significant upgrade over its previous iteration, GPT-3.5.
In a simulated bar exam for U.S. law school graduates, GPT-4 scored within the top 10% of test takers, while GPT-3.5 fell within the bottom 10%.
OpenAI stated that although the two versions may seem similar in casual conversations, “the difference comes out when the complexity of the task reaches a sufficient threshold.”
They also emphasized that “GPT-4 is more reliable, creative, and able to handle much more nuanced instructions.”
A demonstration by Greg Brockman, OpenAI’s president, showcased the AI’s ability to convert a photo of a hand-drawn website mock-up into a functioning website.
Additionally, GPT-4 can assist users in calculating their taxes.
Sam Altman, OpenAI’s CEO, referred to GPT-4 on Twitter as the company’s “most capable and aligned” model with human values and intent, although he admitted that “it is still flawed.”
The new model is 82% less likely to produce disallowed content than its predecessor and achieves 40% higher scores in specific tests of actuality.
Reducing inaccurate responses, or “hallucinations,” has been a significant challenge for many AI systems.
Rishi Jaluria, an analyst at RBC Capital Markets, believes that Microsoft will benefit from the adoption of GPT-4.
The company is not only incorporating OpenAI’s latest technology into its products but also using its Azure cloud to power OpenAI usage.
Jaluria noted that “whenever a company uses this piece of technology,” it will drive workloads through Microsoft Azure, making this development particularly timely given the current economic climate and businesses’ focus on IT spending.
