Meta has become a major player in the artificial intelligence field after a leak exposed its advanced AI language software.
Initially, Meta provided limited access to researchers for developing chatbots similar to OpenAI’s ChatGPT.
However, the leaked code led to the creation of homegrown tools that could potentially rival those offered by tech giants like Microsoft and Google.
Read More: Google starts experimenting with ChatGPT competitors to help boost its AI development
While Meta had previously opened up parts of its AI technology, this particular software was meant for evaluation by researchers and not widespread distribution.
Joelle Pineau, Meta’s vice president of AI research, said: “Someone decided not to respect the terms of use.”
Despite concerns about the misuse of powerful AI, Meta’s software release sparked a surge in AI software development, with numerous positive and exploratory use cases emerging.
The leaked software gained significant popularity.
For example, Stanford University researchers used it to build an AI chatbot at a low cost.
Software company Nomic AI created GPT4All, allowing anyone to run a ChatGPT-style chatbot on their computer.
Meta had been striving to develop its own generative AI software capable of humanlike conversation, similar to OpenAI’s success with ChatGPT and Google’s aims with the Bard chatbot and search tools.
Meta’s development efforts included a fundamental building block called LLaMA (Large Language Model Meta AI), which unexpectedly appeared on the online message board 4chan.
The limited release of LLaMA three months ago led to its public availability due to the leak, igniting grassroots AI development.
Need Career Advice? Get employment skills advice at all levels of your career
Google, a company heavily invested in AI, expressed anxiety over Meta’s software uptake as it could threaten larger companies.
The leak was particularly valuable because Meta had used vast amounts of data and computing power, which small companies and individual developers typically lack.
The code demonstrated how developers could create smaller, less computationally intensive AI products.
However, those attempting to build businesses around Meta’s advanced AI tool encountered a significant intellectual property hurdle.
Meta did not grant permission for commercial use of its code, and there are no plans to change this stance.
Follow us on YouTube, Twitter, LinkedIn, and Facebook.
