AI masters social learning in real-time without pre-collected human data: DeepMind study | Tech News

AI masters social learning in real-time without pre-collected human data: DeepMind study

DeepMind study reveals AI is capable of picking up human social learning skills in real time without prior data. Breakthrough holds potential to reshape AI training and cultural evolution integration.

| Updated on: Dec 11 2023, 15:41 IST
Meta AI is here: Know how to talk to the chatbot on WhatsApp, Instagram, and Facebook
AI social learning skills
1/6 In September 2023, Meta Platforms, the parent company of Facebook, Instagram, and WhatsApp, made a big announcement that it was adding its own native AI chatbot called Meta AI to all of its messaging services. As a result, the chatbot was added to Facebook Messenger, Instagram Direct Messages (DM), and WhatsApp. In September, Meta AI was released in a limited capacity as part of its beta testing, but now it is available to users in the US. (Meta)
AI social learning skills
2/6 Meta AI is an integrated chat experience. You can have an individual conversation with it as well as have it pop up in group conversations to help out with a various things. Meta AI is multimodal, so alongside text responses, it can also generate images from simple text prompts. All you have to type is “@MetaAI /imagine” followed by your prompt. (REUTERS)
AI social learning skills
3/6 What is interesting about Meta AI is that it has also added a bunch of different personalities that you can interact with. These show up as profiles on Facebook Messenger and Instagram. All you have to do is click on them and speak with their AI replicas. Some of the notable people with Meta AI persona include Charli D’Amelio as Coco, Dance enthusiast, MrBeast as Zach, Paris Hilton as Amber, Tom Brady as Bru, and more.  (Meta)
AI social learning skills
4/6 Meta AI can also recommend Reels. This is a new feature which was announced on December 6. Users can go to Meta AI ask it to show Reels for a particular theme and the chatbot will deliver.  In a blog post, the company said, “Say you’re planning a trip to Tokyo with friends in your group chat, you can ask Meta AI to recommend the best places to visit and share Reels of the top sites to help you decide which attractions are must-sees”. (Meta)
AI social learning skills
5/6 Another new feature introduced is called ‘reimagine’ which will be available only in group chats. This is how it works: Meta AI generates and shares the initial image you requested, then your friend can press and hold on the picture to riff on it with a simple text prompt and Meta AI will generate an entirely new image.  (Meta)
image caption
6/6 Apart from this, Meta is working on a number of other features which are currently in testing phase. “We’re exploring ways for you to use AI to help you create the perfect birthday greeting to share with your bestie, edit your own Feed posts, draft a clever introduction for your Facebook Dating profile or even set up a new Group”. Additionally, Meta is also testing a feature that will let users see highlights of the conversations in group chats that they missed out on. (Meta)
AI social learning skills
View all Images
DeepMind study reveals AI is capable of using real-time social learning skills without human data, using observational learning. (Pixabay)

Human intelligence has long relied on cultural evolution and social learning, where knowledge is passed down through imitation of actions and behaviors down through the ages. The question arises: Can artificial intelligence (AI) replicate these social learning skills in real time? Traditionally, AI has employed imitation learning, observing humans and attempting to mimic their actions. However, this method often requires numerous examples and extensive exposure to data. In a groundbreaking study, DeepMind researchers claim that AI agents can exhibit social learning skills in real time without relying on pre-collected human data.

The study focused on observational learning, specifically the imitation of body movements in novel contexts. DeepMind conducted experiments in GoalCycle3D, a simulated environment with varied terrain, footpaths, and obstacles that AI agents navigated. To facilitate learning, reinforcement learning was employed, offering rewards for behaviors leading to the desired result. Also read: Talk to celebrities via AI chatbot! Know how to create your own character on

Mimicking Experts for Accelerated Learning

In the next phase, expert agents, either hard-coded or human-controlled, were introduced. The AI agents quickly learned that mimicking these experts was the most effective way to navigate. Notably, the AI not only learned faster when imitating experts but also applied this knowledge to navigate other virtual paths. Furthermore, the AI agents demonstrated the ability to use their newfound skills even in the absence of experts, showcasing a form of social learning.

Implications for Artificial General Intelligence (AGI)

While acknowledging the need for further research, the authors believe their method could pave the way for cultural evolution to play a crucial role in the development of artificial general intelligence. This breakthrough could revolutionize the AI industry by potentially reducing resource-intensive training and enhancing problem-solving capabilities. Additionally, it prompts the intriguing question of whether AI can assimilate social and cultural aspects of human thought.

In its early stages, DeepMind's discovery opens new avenues for AI development, emphasizing the potential for more efficient and adaptive learning algorithms. The interdisciplinary collaboration between AI and cultural evolutionary psychology could further propel advancements in artificial intelligence.

Follow HT Tech for the latest tech news and reviews , also keep up with us on Whatsapp channel,Twitter, Facebook, Google News, and Instagram. For our latest videos, subscribe to our YouTube channel.

First Published Date: 11 Dec, 15:28 IST