Facebook parent company Meta is ramping up its efforts to advance research about artificial intelligence, as competition between the world's largest tech giants to roll out new AI-powered tools heats up.
On Friday, the social media powerhouse said researchers can apply to access its large language model, LLaMA. Large language models power tools such as Open AI's ChatGPT, which can quickly churn out essays and answer questions. In February, Microsoft launched an AI-powered Bing search engine and Google unveiled an AI chatbot known as Bard.
Meta CEO Mark Zuckerberg wrote in a Facebook post that large language models "have shown a lot of promise in generating text, having conversations, summarizing written materials, and more complicated tasks like solving math theorems or predicting protein structures."
The release of the model shows how AI continues to be an important focus for tech companies as they go head-to-head with each other for users and ad dollars. At the same time, the rising popularity of AI tools that can generate papers, art and other content has also raised ethical concerns about misinformation, plagiarism and bias. Meta says that researchers often have limited access to large language models, hampering work that aims to address these problems.
"We believe that the entire AI community -- academic researchers, civil society, policymakers, and industry -- must work together to develop clear guidelines around responsible AI in general and responsible large language models in particular," Meta said in a blog post. "We look forward to seeing what the community can learn -- and eventually build -- using LLaMA."
Though access to LLaMA is limited to researchers Meta approves, advancing AI overall could benefit the social media giant along with other tech companies. In February, Zuckerberg said Meta has been "deploying AI to help engineers be more productive" after layoffs and restructuring at the company. Facebook also uses AI to recommend content that people might be interested in, and it applies AI to the company's ad business.
"Generative AI is an extremely exciting new area with many different applications, and one of my goals for Meta is to build on our research to become a leader in generative AI in addition to our leading work in recommendation AI," Zuckerberg said during a call with analysts.
Popular AI-powered applications such as ChatGPT and image generator Dall-E have fueled the AI arms race, but tech companies including Meta have been working to advance AI for years. Meta has also seen firsthand what problems come with relying on automated technology.
In 2022, Meta released a live demo of a chatbot called BlenderBot 3, but it wasn't long before the bot spewed out conspiracy theories, antisemitic remarks and even insulted Facebook. In November 2022, Meta also paused a demo of a language model called Galactica that could summarize academic research, solve math problems and create other science content. The AI tool had generated misinformation.
Editors' note: CNET is using an AI engine to create some personal finance explainers that are edited and fact-checked by our editors. For more, see this post.