Meta Takes Generative AI Plunge, Google Enhances Encryption, More News
It was only a matter of time before Meta jumped on the ChatGPT hype train, so this week's announcement that the Menlo Park, Calif.-based company is creating a new “top-level group” focused on the development of generative AI and its possible application to the metaverse comes as no surprise.
According to CEO Mark Zuckerberg, the new group is tasked with “building delightful experiences" around this technology (generative AI) and pushing it into Meta's different products, with a particular focus on the metaverse.
Zuckerberg remains committed to metaverse development despite the massive losses the company racked up in 2022 following its rebranding of the company.
While it is early days in the development of the metaverse — just as it is in the world of generative AI — Zuckerberg's vision for this new team in practical terms remains unclear. In a Facebook post announcing its creation, he wrote that it would be focused on building creative and expressive tools.
The long-term ambition, he wrote, is to develop AI personas that will be able to help people in all kinds of ways.
He added that the company is looking at new experiences with text “like chat in WhatsApp and Messenger,” with “images (like creative Instagram filters and ad formats)," and with "video and multi-modal experiences."
However he admitted a great deal of foundational work would be needed before the emergence of any of these “really futuristic experiences.”
Zuckerberg's post received a high number of responses as might be expected (14,000 comments at the time of writing). But it leaves us no closer to knowing exactly what kind of a team he is talking about and how "top-level" it will actually be, nor how it would work with Meta's existing AI experts such as chief AI scientist, Yann LeCun. The reticence could possibly just be caution after being so badly burnt on his metaverse ambitions last year.
In the earnings report published at the end of January, Meta indicated that its Reality Labs division, which develops the company’s virtual reality technologies and projects, posted a $4.28 billion operating loss in the fourth quarter, bringing the total for 2022 to $13.72 billion.
The impact on the company has been enormous with 11,000 people laid off in November alone and a “year of efficiency” in the offing.
Microsoft’s Kosmos-1 Moves Closer to Artificial General Intelligence
Meta isn't the only company pursuing AI ambitions. While ChatGPT has captured the media (and users') attention, with great predictions around its future (has anyone queried ChatGPT about what it believes its future will be?), it might be easy to view it and Microsoft’s investment in it as the last word in generative AI.
And while Microsoft is optimistic about its possible use in the digital workplace, it is not limiting itself to OpenAI. This week, for example, researchers at Microsoft unveiled Kosmos-1, which could have huge potential for the digital workplace and beyond.
Kosmos-1, according to a paper published on arxiv.org, is a multimodal language model which, the researchers say, can analyze images for content, offers visual text recognitions and understands natural language structures, among other things.
Large language models have become one of the new currencies in AI, the best known of which is OpenAI’s GPT (Generative Pre-trained Transformer) model.
Kosmos-1 is a step beyond that. As a multimodal large language model (MLLM) it can respond to language prompts and visual ones too. The uses in the workplace would be multiple.
The paper, "Language Is Not All You Need: Aligning Perception with Language Models," describes how the MLLMs are another step towards creating artificial general intelligence (AGI), which is what many people are hoping GPT-4 from OpenAI will offer when it is finally released — in spite of warnings to the contrary from OpenAI CEO Sam Altman.
Artificial general intelligence (AGI) describes the ability of an intelligent agent to understand or learn any intellectual task that humans are capable of.
In the paper, Kronos-1 responded to a number of questions that would be beyond the capacity of current AI models to respond to.
The most striking of these images shows a picture of a cat in front of which someone is holding a piece of paper with a drawn smile on it. The input question reads: Explain why this photo is funny? The output response reads: The cat is wearing a mask that gives the cat a smile (page 2 of the paper). The paper include many other examples.
The researchers say the model is at an early stage, but are confident that they will be able to scale it up from visual prompts to verbal prompts too.
The result for workers would be an AI tool that could understand any workplace media and act on it. If it is scaled-up to respond to speech, then it would also be able to respond to any verbal communications in the workplace.
How to Future-Proof Your Employee Experience Strategy in 2023
A framework to navigate through economic uncertainty
The Essential Role of Communicators in Fostering Wellbeing in the Digital Workplace
Join us for practical insights on how digital communicators can support employees to thrive in the digital workplace
Addressing Employee Needs and Wants with a Digital Workplace
The workplace is getting more and more digital – both in how we work and where we work
Maintaining a Human-Centered Approach During Digital Transformation
When it comes to digital transformation - people drive change, not technology
The Evolution of Employee Recognition
Leveraging the power of appreciation to improve the employee experience
How to Build a More Innovative and Resilient Workplace Culture
What would happen if every member of your team came to work focused on finding solutions and creating better results?
Google Gmail and Calendar Get Enhanced Encryption
Security remains a big issue in the digital workplace, particularly securing apps that use and share company data. For better or for worse, email remains one of the most used apps to share data and Google has spent a great deal of time and development dollars securing Gmail and Calendar.
This week, the company made good on a promise made last December, and announced general availability of client-side encryption (CSE) for both apps.
Google's Ganesh Chilakapati and Andy Wen explained the new encryption capabilities give users sole control over their encryption keys and, consequently, sole control over access to their data. With CSE, users can send emails or create events with internal and external collaborators that have been encrypted before they reach the Google servers.
CSE is now available across the entire Workspace suite, which makes it attractive to all kinds of workplaces, particularly in highly regulated sectors. It also reduces compliance problems for enterprises and public sector organizations as neither Google, not to mention foreign governments, can access the data.
As of the beginning of March CSE is available globally to Workspace Enterprise Plus, Education Standard and Education Plus customers.
Robin Improves Hybrid Workplace Management
Boston-based hybrid workplace experience platform Robin has added a number of new features designed to give organization leaders and employees insights into the creation and implementation of hybrid workplace policies.
Citing LinkedIn data, the company noted the drop in employees working in fully remote models — from 20% in April 2022 to 14% in October 2022 — with more of them adopting hybrid work models.
Organizations operating with hybrid models have faced major challenges in managing them, especially with the emergence of three different kinds of work:
- Fully flexible: Employees can come and go as they choose.
- Structured hybrid: Employees are required to be in the office based on parameters set by the employer.
- On-site: Workers are located on-site on a permanent basis.
The new additions enable administrators to set the company’s hybrid work policy, including any combination of days a week onsite or in the office. From the employee perspective, it also becomes easier to manage, as the number of days in and out of the office appears on the homepage giving them an overview of whether they are meeting targets or not.
Robin was founded in 2014 and, according to Crunchbase has raised $59.1 million to date.
Hexa Raises $20.5M for AR, VR Development
Finally this week, Tel Aviv, Israel-based Hexa, a 3D asset visualization and management platform, announced it has raised a $20.5 million funding in a Series A round.
According to a company statement, its technology offers organizations the capabilities of digitizing products using existing 2D images and AI, and then creating a new 3D model, a digital twin of the physical object.
The digital twin, the statement says, can be used on websites, social media and even in AR offerings.
The company was founded in 2018 and has raised $27.2 million to date.
Have a tip to share with our editorial team? Drop us a line:
About the Author
David is a full-time journalist based in Ireland. A partisan of ‘green’ living and conservation, he is particularly interested in information management and how enterprise content management, analytics, big data and cloud computing impact on it.