AI news
Meta to release commercial AI model in effort to catch rivals 🤖
Meta, previously known as Facebook, is preparing to release a commercial version of its artificial intelligence (AI) model, which will allow businesses and start-ups to build custom software on top of the technology. This move is seen as a way for Meta to compete with Microsoft-backed OpenAI and Google, who are currently leading in the development of generative AI.
Meta's AI model, known as LLaMA, was initially released to researchers and academics, but the new version will be more widely available and customizable by companies. Unlike its competitors, Meta's AI models are open-source, meaning the details of the new model will be publicly released. This is in contrast to models like OpenAI's GPT-4, which are considered "black boxes" where the data and code used to build the model are not available to third parties.
Meta's approach to AI is seen as a way to change the competitive landscape of AI, with open-source platforms that are as good as those that are not. This release is part of a race among tech groups to establish themselves as dominant AI participants.
While Meta's technology is open-source and currently free, the company has been exploring charging enterprise customers for the ability to fine-tune the model to their needs using their own proprietary data. However, there are no current plans to charge for this in the upcoming release.
The company has been investing heavily in AI research and development for over a decade. In 2021, CEO Mark Zuckerberg announced a pivot to build a digital world known as a metaverse, spending more than $10bn a year on the project. As part of this, the company has hinted at creating multiple AI chatbots for individuals, advertisers, and businesses across Meta platforms Instagram, WhatsApp, and Facebook, powered by its large language models (LLMs).
The benefits of open-source models include a higher take-up by users who then input more data for the AI to process, making the AI more powerful. Open-source models also allow researchers and developers to spot and address bugs, improving the technology and security simultaneously.
However, there are risks with open-source AI, including potential misuse by bad actors and legal risks concerning intellectual property and copyright. Despite these risks, Meta and other AI companies see the potential in releasing open-source versions of their technology.
ft.com
New features
Not camera ready? Go virtual. Your Meta avatar now works in Messenger and Instagram video calls📺
The article from Mashable discusses a new feature from Meta that allows users to replace their real faces with their Meta avatars during Messenger or Instagram video calls. These avatars are virtual, cartoon-like versions of the user that can be created in the Facebook or Instagram account settings. Previously, these avatars were used as stickers on Facebook, Instagram, or Messenger.
The new feature allows the avatars to be used as an animated version of the user in a video call, primarily aimed for those times when users are "not camera-ready". The Messenger and Instagram apps will use the phone's video camera to replicate the user's facial expressions on the avatar's face. This feature works on both Android and iOS.
In addition to this, the company announced other avatar-related features, such as animated avatar stickers for Instagram, Facebook Stories and Reels, and one-on-one message threads on Messenger and Instagram. A new option allows users to create their avatar on Facebook and WhatsApp by taking a live selfie and having the app generate the avatar, although this feature is currently being tested with a limited audience.
mashable.com
Tools
Meta’s First Responsible Business Practices Report 🔖
Meta has released its first Responsible Business Practices Report, which outlines the company's impact on society and its approach to responsible operation. The report focuses on four pillars:
Operating transparently and sustainably: Meta aims to conduct business responsibly, ethically, and transparently to maintain stakeholder trust and its global operating license. Empowering people and partners: The company is committed to treating employees, suppliers, and other partners with respect and dignity, implementing high safety and care standards, and ensuring that supply chain partners do the same. Building responsibly: Meta strives to design, develop, and improve its products and platforms to enable safe and inclusive experiences that improve people's lives. Unlocking potential for good: The company aims to empower people worldwide to share ideas, offer support, and make a difference through the use of its products. The report includes a comprehensive data index prepared in accordance with various international standards and initiatives. Highlights include the launch of Family Center and Privacy Center, spending over $1 billion with diverse suppliers, raising over $7 billion for social causes through its apps, and contracting over 10,000 MW of renewable energy projects.
The report also includes data on Meta's diversity, equity, and inclusion efforts, replacing a previously standalone DEI report. Despite reducing its workforce and hiring rates, Meta remains focused on growing and retaining a diverse workforce. The company will continue to evolve its responsible business approach based on feedback and results to contribute positively to society. The full report is available on the Meta website for more details.
fb.com
Social Mind: Social and AI Insights
Welcome to Thursday's update on the latest in the world of social media 🙌
Today I have a number of meetings plus the usual content creation.
Tonight I'm looking forward to All Talks Running, an event hosted in Cardiff by elite runner Olivia Tsim - should be very interesting.
Have a great day!
JP
Chat with me on Threads📞Call/WhatsApp on 029 2078 9139