Meta is on the brink of launching a commercial version of its artificial intelligence (AI) model, which will enable start-ups and businesses to build customized software using this technology.
This move positions Meta as a competitor to Microsoft-backed OpenAI and Google, both of which are leading the race in developing generative AI. Meta’s software, powered by large language models (LLMs), can generate text, images, and code. These models are trained on massive amounts of data and require significant computing power.
Earlier this year, Meta released its own language model called LLaMA to researchers and educators. However, the upcoming version will be more widely accessible and customizable by businesses, according to three insiders. The release is expected in the near future, confirms one of the sources.
Meta describes its LLMs as “open-source,” meaning that the details of the new model will be made publicly available. This is in contrast to competitors like OpenAI, whose latest model, GPT-4, is a “black box” where the data and code behind the model are not shared with third parties.
Speaking at a conference in Aix-en-Provence, Meta’s Vice President and Chief AI Scientist, Yann LeCun, said, “The competitive landscape of AI is going to completely change in the coming months, maybe even weeks, with the emergence of open-source platforms that are on par with the non-open ones.”
Meta’s imminent release coincides with a race among major Silicon Valley tech companies to establish dominance in the field of AI. In an article for the Financial Times, Meta’s Global Affairs Chief, Nick Clegg, praised the merits of an open-source approach, stating that “openness is the best solution to alleviate AI-related concerns.” Furthermore, adopting an open model allows Meta to catch up with its competitors, as it enables companies of all sizes to improve the technology and build applications on top of it.
Meta has been actively involved in AI research and development for over a decade. However, it has seemed to be lagging behind since the release of OpenAI’s ChatGPT last November, which prompted other tech giants to launch similar products.
According to one individual familiar with Meta’s high-level strategy, their goal is to reduce OpenAI’s current dominance in the AI market.
Regarding pricing, although Meta’s technology is currently free and open-source, insiders suggest that the company has been considering charging enterprise customers for the ability to adapt the model to their specific needs using proprietary data. However, no current plans to implement this fee have been revealed, and it will not be part of the upcoming release.
Joelle Pineau, Meta’s Vice President of AI Research, declined to comment on the development of a new AI model and its possible monetization. She did mention that “releasing something as open-source does not mean completely giving up intellectual property rights.”
In 2021, Meta’s CEO, Mark Zuckerberg, announced a shift towards building a metaverse, a digital world filled with avatars. The company has been investing over $10 billion annually in this project, which has faced criticism from investors. Consequently, Meta has recently accelerated its AI investment.
Meta established a generative AI unit earlier this year, led by Chief Product Officer Chris Cox. Pineau explained that Cox’s team is involved not only in AI research but also in product development, including the creation of entirely new businesses.
Zuckerberg and other executives have hinted at developing multiple AI chatbots for individuals, advertisers, and businesses across Meta’s platforms, including Instagram, WhatsApp, and Facebook. These chatbots would be powered by Meta’s LLMs.
One advantage of open-source models is their ability to attract more users, who contribute more data for AI processing. The more data an LLM receives, the more powerful its capabilities become.
Moreover, open-source models enable researchers and developers to identify and address bugs, leading to improvements in both technology and security. This is crucial, as technology companies like Meta have faced years of scrutiny over privacy and misinformation scandals.
Although offering free software may seemingly contradict profit-making objectives, experts believe that this strategy can help corporations tap into new markets. One individual familiar with Meta’s perspective explained, “Meta realized they were falling behind in the current AI hype cycle. This open-source approach allows them to appear charitable and committed to giving back to the community.”
Nevertheless, open-source AI also carries risks of misuse and abuse by malicious actors. Notably, child safety groups have reported an increase in AI-generated child sexual abuse imagery online.
There are also concerns regarding intellectual property and copyright in the context of open-source AI. Comedian and actor Sarah Silverman recently filed a lawsuit against Meta and OpenAI, alleging that her work was used to train models without her consent.
Meta released its open-source model LLaMA to researchers in February. However, it later leaked more widely through the online forum 4chan, leading developers to use it in commercial products in violation of Meta’s licensing rules.
Joelle Pineau acknowledged that this unauthorized distribution presents challenges, stating, “This model is out there in ways that we wish it wasn’t.”
Other AI companies, including French start-up Mistral, are also exploring the possibility of releasing open-source versions of their technology. OpenAI is considering developing an open-source LLM, provided they can minimize misuse risks to an acceptable level.
Meta’s AI Chief, Yann LeCun, highlighted the choice between locking away AI technology or encouraging contributions from a global community through open-source platforms: “We have a choice between deciding that artificial intelligence is too dangerous a technology to remain open and putting it under lock and key and in the hands of a small number of companies that will control it, or, on the contrary, open-source platforms that call for contributions… from all over the world.”
Additional reporting by Tim Bradshaw
Denial of responsibility! VigourTimes is an automatic aggregator of Global media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, and all materials to their authors. For any complaint, please reach us at – [email protected]. We will take necessary action within 24 hours.
