Artificial intelligence

Meta to Release Trading AI Model in Aim to Catch Rivals

Meta is set to release a commercial version of its AI model, allowing start-ups and enterprises to build custom software on top of technology.

This decision will allow Meta to compete with Microsoft-backed OpenAI and Google, which are advancing in the race to develop generative solutions. AI. The software, which can create text, images and code, is powered by large language models (LLMs) which are trained on huge amounts of data and require high computing power.

Meta released its own language model, known as LLaMA, to researchers and academics earlier this year, but the new version will be more widely available and customizable by businesses, three people familiar with the plans said. The release is expected imminently, one of the people said.

Meta says its LLMs are “open-source,” meaning details of the new model will be made public. This contrasts with the approach of competitors like OpenAI, whose latest GPT-4 model is a so-called black box in which the data and code used to build the model are not available to third parties.

“The competitive landscape for AI is going to completely change in the coming months, maybe weeks, when there will be open source platforms that are actually as good as those that aren’t.” , vice president and chief scientist of AI at Meta, said Yann LeCun during a conference in Aix-en-Provence last Saturday.

Meta’s impending release comes as a race among Silicon Valley tech groups to establish themselves as dominant AI participants heats up.

Write in the Financial Times this week, Meta’s head of global affairs, Nick Clegg, extolled the virtues of an open source approach, saying “openness is the best antidote to fears surrounding AI.” But the move also helps Meta in its attempts to catch up with rivals, as an open model would allow businesses of all sizes to improve on the technology and build apps on it.

Meta has been working on AI research and development for more than a decade, but appears to be on the back foot after the release of OpenAI’s ChatGPT, a conversational chatbot, in November, prompting other Big Tech groups to launch similar products.

“The goal is to diminish OpenAI’s current dominance,” said a person with knowledge of high-level strategy at Meta.

Meta declined to comment.

Although Meta’s technology is open source and currently free, two people familiar with the matter said the company plans to charge enterprise customers for the ability to tune the model to their needs using their own proprietary data. One person said that there are no current plans to charge and that Meta won’t in the next release. Meta’s intention to release its AI model under a commercial license was first reported by The Information.

Joelle Pineau, vice president of AI research at Meta, declined to comment on the development of a new AI model and how it might be monetized, but said, “Ultimately, because you post something [open source]you do not completely relinquish the intellectual property of this work.

“We did not hesitate to use these models [in our] products,” she added.

In 2021, Chief Executive Mark Zuckerberg announced a pivot to build a digital world filled with avatars known as the Metaverse and has spent over $10 billion a year on the project. This costly ambition has proven unpopular with investors and Meta recently raced to increase its investment in AI.

Earlier this year, the social media giant set up a generative AI unit led by product manager Chris Cox. Pineau said Cox’s team straddles the research side of AI but also product development, as it “creates totally new businesses.”

Zuckerberg and other executives hinted at a push to create more AI chatbots for individuals, advertisers and businesses on the Instagram, WhatsApp and Facebook Meta platforms, powered by his LLMs.

The benefit of open source models includes higher adoption by users who then enter more data for the AI ​​to process. The more data an LLM has, the more powerful its capabilities can become.

Additionally, open-source models allow researchers and developers to spot and fix bugs, simultaneously improving technology and security — at a time when tech companies such as Meta have faced years of scrutiny from various privacy and disinformation scandals.

Although providing software for free may seem contrary to making money, experts believe that companies can also use this strategy to conquer new markets.

“Meta realized that they were behind the current AI hype cycle, and that gives them a way to open up the ecosystem and make it look like they’re doing what they’re doing. it takes, by being charitable and giving back to the community,” said a person familiar with the company’s thinking.

Yet open-source AI presents obvious risks, which can be shaped and abused by bad actors. Child safety groups are reporting an increase in AI-generated child sexual abuse images online, for example.

The researchers also found that a previous Meta AI model, BlenderBot 2, released in 2021, was spreading misinformation. Meta said this makes the BlenderBot 3 more resistant to this content, although users have always found it to generate false information.

There are also regulatory and legal risks regarding intellectual property and copyright. On Monday, comedian and actress Sarah Silverman filed a lawsuit against Meta and OpenAI for claiming her work was used to train models without her consent.

Meta released its open source LLaMA model to researchers in February. A month later, it was leaked more widely via online forum 4chan, prompting developers to rely on it in violation of Meta’s licensing rules, which specify that it must not be used in commercial products.

“That model is out there in a way that we wish it weren’t,” Pineau said.

Other AI companies, such as a French start-up Mistral, are also exploring the possibility of releasing open source versions of their technology. OpenAI, which has previously published open-source AI models for speech and image recognition, said its team plans to develop an open-source LLM, provided it can reduce the risk of misuse below d a minimum threshold.

“We have a choice between deciding that artificial intelligence is too dangerous a technology to remain open and locking it up and in the hands of a small number of companies that will control it,” said LeCun, head of the Meta AI. “Or, on the contrary, open source platforms that call for contributions. . . From all over the world.”

Additional reporting by Tim Bradshaw

Leave a Reply