Meta Confirms Use of Data from Facebook for Training AI

Meta acknowledges the utilization of data from Facebook and Instagram to train an ambitious Artificial Intelligence (AI). Let's explore this development and the crucial questions surrounding personal privacy and copyright in the rapidly evolving digital landscape.

Meta, the tech giant formerly known as Facebook, has announced the use of data from Facebook and Instagram to train an ambitious Artificial Intelligence (AI) model. Let's delve into the details of this development and the crucial questions surrounding personal privacy and copyright in the rapidly evolving digital landscape.

Meta acknowledges the utilization of data from Facebook and Instagram to train an ambitious Artificial Intelligence (AI). Let's explore this development and the crucial questions surrounding personal privacy and copyright in the rapidly evolving digital landscape.

The Birth of Multi-Personality AI Chatbot

Meta, the world-class technology conglomerate, has officially acknowledged its use of content from Facebook and Instagram to train an entirely new Artificial Intelligence (AI) model. This move has stirred controversy and raised questions surrounding user privacy and security.

On September 27th, Meta officially introduced the "multi-personality" AI chatbot. Nick Clegg, Meta's Vice President of Global Affairs, announced in an interview with Reuters that the training data for this AI is sourced from public posts and images shared by users on the Facebook and Instagram platforms. However, Clegg emphasized that the AI does not utilize information from private posts or content shared exclusively with friends, family, or in private messages.

Representatives from Meta have stated that they have implemented various technical measures to control which data the AI can access and utilize. However, they have not disclosed specific details about how this model operates. Clegg mentioned, "We are working to exclude datasets containing excessive personal information. Meta also refrains from using data from websites like LinkedIn for AI training due to privacy concerns."

This new AI, described by Meta CEO Mark Zuckerberg as "Meta's AI Assistant capable of conversing like a human," is built upon the Llama 2 language model and a text-to-image conversion model called Emu. The product is capable of generating text, audio, and images while also having real-time access to information through collaboration with Microsoft's Bing search engine.

The AI model includes 28 chatbots created based on famous personalities such as Snoop Dogg, Tom Brady, Kendall Jenner, and Naomi Osaka. However, during the initial testing phase, Meta AI faced considerable controversy. According to the Wall Street Journal, adding personality traits to chatbots has led to harmful behavior and excessive responses. Some chatbots tend to provide racially biased information or delve too deeply into users' personal lives. Many Meta employees even declared their reluctance to use the new chatbots due to the peculiar content generated by AI.

Chatbots and AI Models: Controversial Emergence of Personalization. 

One of the significant concerns raised by the community as Meta introduces highly customizable AI models based on user data is related to copyright issues. According to Reuters, copyright matters concerning AI remain a subject of debate. Some companies allow the easy creation of iconic images like Mickey Mouse, while others have paid for copyrights or avoided using them in AI training to prevent intellectual property disputes.

For example, OpenAI has signed a six-year contract with Shutterstock to use the company's library of images, videos, and music for AI training purposes. When asked whether Meta has taken any measures to avoid copying copyrighted images, Meta's spokesperson emphasized the terms of service prohibiting users from creating content that violates privacy or intellectual property rights.

This marks a significant milestone in the utilization of user data for AI training and raises essential questions about privacy and copyright in an increasingly digital world.

Source: Reuters, Business Insider

Read more: 

- Top 5 .NET Development Trends To Watch In 2023 

- ChatGPT And Bing AI: Which Is The Superior AI Chatbot

- Improving Order Management Through Mobile App Development

We hope you can find the best partner for your business. Take a look at our portfolio to see the full range of projects our professionals have tackled. To discuss your idea — leave us a message via:

Email: [email protected]
Hotline: +84 28 71092939   |  +84 933 848989
Website: www.aegona.com
Fanpage: Aegona Ltd
Address: Orbital Building (QTSC Building 9), Quang Trung Software City, District 12, Ho Chi Minh City, Viet Nam