META has unveiled its most powerful AI model yet – and CEO Mark Zuckerberg claims it is on track to beat out its fiercest competitor.
The tech behemoth has released Llama 3.1, a free, open-source AI model, after months of buildup.
Llama 3.1 is significantly more advanced than the smaller Llama 3 models unleashed just a few months ago.
It is so advanced that it outperforms OpenAI’s GPT-4o across several benchmarks, including those designed to test reasoning.
Zuckerberg made the formal announcement in a letter published today on Meta’s company blog.
The tycoon also made the case for an open-source AI model, which companies can train on custom data and fine-tune to their liking.
Meta has partnered with industry leaders like Microsoft, Google, and Nvidia to kickstart this process.
“People often ask if I’m worried about giving up a technical advantage by open sourcing Llama, but I think this misses the big picture,” Zuckerberg wrote.
Releasing Llama 3.1 to the world will help it “develop into a full ecosystem of tools, efficiency improvements, silicon optimizations, and other integrations,” he continued.
To drive the point home, Zuckerberg referenced the company’s “long history of open source projects and successes.”
Meta “saved billions of dollars” by making its server, network, and data center designs public.
“This approach has consistently worked for us when we stick with it over the long term,” Zuckerberg professed.
Unlike its competitors, Meta‘s business model does not rely on selling access to AI models – meaning open releases don’t undercut “revenue, sustainability, or ability to invest in research.”
The company hasn’t disclosed where it sourced the data to train Llama 3.1, but it previously came under fire for scraping information from Instagram and Facebook.
Meta mentions training the model on synthetic, or AI-generated, data in other blog posts.
Zuckerberg closed out the letter addressing concerns about AI, which he broke into two categories: unintentional and intentional harm.
Examples of unintentional harm are “bad health advice” or the futuristic worry that models may “unintentionally self-replicate or hyper-optimize goals to the detriment of humanity.”
Intentional harm, meanwhile, can be summed up as the work of a “bad actor.”
As most “concerns people have around AI” fall into the unintentional harm category, this is why open-source software is the best way forward, Zuckerberg continued.
To mitigate risk, Llama 3.1 comes equipped with features like Llama Guard, which is intended to safeguard human-AI conversations.
Zuckerberg is optimistic for the future.
In a video posted to Instagram, the tech tycoon claimed Meta AI was on track to become “the most-used AI assistant in the world by the end of the year.”
That record is currently held by OpenAI‘s ChatGPT, which boasts over 100 million active users.
Despite the fierce competition, Meta aims to establish itself as a pioneer, implementing a range of AI features across platforms and service.
The company’s flagship model has received some updates as well, with Meta deeming it “more creative and smarter” than ever before.
Artificial intelligence is a highly contested issue, and it seems everyone has a stance on it. Here are some common arguments against it:
Loss of jobs – Some industry experts argue that AI will create new niches in the job market, and as some roles are eliminated, others will appear. However, many artists and writers insist the argument is ethical, as generative AI tools are being trained on their work and wouldn’t function otherwise.
Ethics – When AI is trained on a dataset, much of the content is taken from the Internet. This is almost always, if not exclusively, done without notifying the people whose work is being taken.
Privacy – Content from personal social media accounts may be fed to language models to train them. Concerns have cropped up as Meta unveils its AI assistants across platforms like Facebook and Instagram. There have been legal challenges to this: in 2016, legislation was created to protect personal data in the EU, and similar laws are in the works in the United States.
Misinformation – As AI tools pulls information from the Internet, they may take things out of context or suffer hallucinations that produce nonsensical answers. Tools like Copilot on Bing and Google’s generative AI in search are always at risk of getting things wrong. Some critics argue this could have lethal effects – such as AI prescribing the wrong health information.
The Meta AI assistant is available in more languages including French, German, and Hindi.
There’s also the “Imagine Me” feature, which scans a user’s face through their phone camera and inserts their likeness into AI-generated imagery.
The tool is available in beta and can be invoked by typing “Imagine me” followed by any safe-for-work activity.
An experimental version of Meta AI is rolling out to the Quest headset next month, beginning in the United States and Canada.
Users will be able to ask simple questions, translate text, and get real-time information through Bing integration.
