Madrid, July 19 (EFE).- US technology giant Meta, the parent company of social media apps such as Facebook, Instagram and WhatsApp, believes that artificial intelligence (AI) is a “ubiquitous” technology that will affect how “everybody interacts with the online world in the years ahead.”
“The metaverse does not work without AI, they are umbilically linked. One does not work without the other. You can’t build a virtual world, you can’t build augmented reality, you can’t build mixed reality without AI,” Nick Clegg, former UK deputy prime minister and now president of global affairs at Meta, tells EFE in an interview via videoconference.
“AI is already integrated into the product experience. It means we will be able to move further and faster in building useful applications for and useful content and more user friendly content for people who want to experience augmented and virtual reality,” he stresses.
On Tuesday, Meta, together with Microsoft, launched a new open source model for developers based on AI and predictive language called Llama 2, which Clegg calls “the most sophisticated language model in the world.”
Llama uses sequences of words to generate predictive text with content from more than 27 languages and with more than 70 billion parameters.
Open source leaves the ball in the companies’ court: “It all depends now on how innovators, developers, entrepreneurs, academics and researchers choose to use these models. That’s the magic of open sourcing. You don’t know exactly how people are going to use it,” he says.
Llama was presented in February of this year and its development has involved around 350 people carrying out tests “to make sure that it can be used responsibly and safely”, he points out.
For the Meta executive, it can be applied in hundreds of ways, from company chatbots (tools in which AI has “conversations” with users), to helping people write a wedding speech or for analyzing data and generating information.
“It’s going to boost innovation. It will make it easier, for instance, for banks to detect fraud through AI or for medical diagnostics to diagnose the data from MRI scans and X-rays and so on,” he says.
However, he does not rule out inappropriate use of the technology: “Users can try to generate disinformation, fake news and other types of content more rapidly than before,” he points out.
To combat that type of usage, the company led by Mark Zuckerberg also uses AI to tackle deep-fakes, fake news and disinformation “regardless of whether it has been produced by a human being or a machine” Clegg says, pointing out that AI is actually “one of the most effective tools we have” in combating that kind of activity.
In this regard, he highlights how they have managed to bring hate speech content on Facebook down to just 0.02%, a drop of more than 50% in the last 18 to 24 months.
“Generative AI (such as Llama2) is an even more powerful tool for us to stop the distribution of that type of content regardless of who it has been generated by,” he stresses.
As for the future, Clegg believes there will be two areas to watch: messaging apps and augmented and virtual reality, where he predicts “constant improvements”.
“Messaging apps will be increasingly used in more versatile ways. In India more than 400 million people use WhatsApp in a more versatile way than we do in Spain or the UK,” he concludes. EFE