Privacy experts are concerned about Facebook’s recent AI developments, alerting users to possible data breaches and advising them to be mindful of their digital footprints.
Meta’s CEO has sparked serious concerns after discussing the company’s intentions to leverage user data from Instagram and Facebook for training. He claims that Meta has a plethora of data that is even more than what was previously observed for training ChatGPT. Consequently, the company will use this data to train its artificial intelligence (AI) system.
Using posts and comments for chatbot training
The tech giant aims to use posts and comments to train chatbots, and this has raised serious concerns. Several people are concerned about the toxicity as well as privacy.
Just like OpenAI came under fire for scraping up the copyrighted data of artists and writers to train its models, Facebook stands to face reputational blowback for revealing plants to exploit people’s data all over again.
‘In the same way OpenAI has come under fire for scraping up the copyrighted data of artists and writers to train its models, Facebook stands to face reputational blowback for exploiting people’s data all over again.’ https://t.co/kgmLNSvfHe
— Jesse Felder (@jessefelder) February 6, 2024
Meta’s Facebook app is a major deal, as its user database keeps increasing even at this moment. The firm is said to be on a roll as per the latest stats.
However, this continuous growth does not have Mark Zuckerberg celebrating just yet. Mark hopes to leverage this growth by making use of Facebook and Instagram data to produce content that’s not unique but powerful. Facebook is the internet for many people.
Furthermore, the CEO expresses hope that using data and feedback to the fullest extent possible for a multitude of items will be the next essential component of their lineup.
Using photographs and other datasets
Although there are a lot of photographs shared on these apps, they estimate that the number is greater than the typical datasets of public texts and posts made on other platforms.
The company may be able to develop a larger and more intelligent chatbot than what is now in demand if it has data that exceeds what was observed throughout a broad range of text used for ChatGPT training.
According to Bloomberg, Meta may be well-positioned for two main reasons. To begin with, it is related to the substantial amount of data involved. Second, the conversational nature of the majority of the data implies that it may be leveraged to give the chatbot an advantage over the competition.
Mark reportedly sent a message to his employees that the next key part of their playbook is learning from unique data and feedback loops in their products. On Facebook and Instagram, there are hundreds of billions of publicly shared images and tens of billions of public videos, which they estimate is greater than the Common Crawl dataset, and people share large numbers of public text posts in comments across their services as well.
Concerns from users
However, there are two major warning signs in play here. For example, Meta is using user-deemed personal posts to train AI. Meanwhile, another important aspect to note that has sparked grave concerns about privacy being violated is conversations between friends or loved ones that happen in the comments section.
Significantly, the volume of data arising in the comments section was toxic.
Zuckerberg’s Secret Weapon for #AI Is Your Facebook Data.
Perhaps we shouldn’t trust a billionaire who’s been reckless with personal information that he wants to exploit yet again.
By @parmy https://t.co/9ENlJsjeyE via @opinion
— alain servais (@aservais1) February 6, 2024
Several users are sharing insights, some with personal attacks on the Meta CEO. This can be deemed worrisome and racist by many.
According to some X (formerly Twitter) users, they shouldn’t trust a billionaire who’s been reckless with personal information. They also claim that he wants to exploit him yet again.