Connect with us

Tech

Facebook’s Artificial Intelligence Errors Questioned by Critics

Avatar of CTN News

Published

on

Facebook's Artificial Intelligence Errors Questioned by Critics

Facebook, which rebranded itself, Meta, is faced with the daily challenge of balancing supporting free expression while blocking out unwanted material like images of child sexual abuse, violent incitement, and financial scams.

However, Facebook’s artificial intelligence programs that are designed to block and filter out unwanted material are prone to making mistakes and banning users over commonly used words or names.

The social network’s automated system flagged discussions about a common backyard tool as inappropriate sexual talk in a group for gardeners and has even blocked Facebook sign in attempts.

Several years ago, Facebook froze the accounts of Native Americans because its computers mistakenly believed that names like Lance Browneyes were fake.
It repeatedly rejected advertising from businesses selling clothing for people with disabilities, mostly because it confused the product with medical promotions, which are prohibited.

People, businesses, and groups serving the public interest, such as news organisations, suffer when Facebook cuts off their accounts and they cannot find help or figure out what went wrong.

According to the Wall Street Journal, Facebook may make 200,000 mistakes every day.

Facebook essential to companies

Several social network researchers told the Wall Street Journal that Facebook and its peers could do more to make fewer mistakes and mitigate the damage when they do.

There is also a broader question: Are we OK with companies that are so essential that when they make mistakes, we can’t do much?

Facebook has been criticized for not making it easy for users who had their posts deleted, or whose accounts were disabled, to see what rules they broke and appeal decisions made by the company.

The semi-independent Facebook Oversight Board has also said the company needs to make it easier for users to understand what rules they broke, and appeal decisions made. However, Meta has done too little to help.

The researchers want to analyze Facebook’s data to see how it arrives at decisions and how often it makes errors. Facebook opposes that idea saying it’s an invasion of user privacy. Yet Facebook sells users’ data to the highest bidder.

According to Meta, the company is working to be more transparent and spends billions of dollars on computer systems and people to supervise communication in its apps.

Users will always disagree with the company’s decisions regarding posts. Critics, however, argue that the company has not done enough.

This article originally appeared in The New York Times.

 

Continue Reading

CTN News App

CTN News App

Recent News

BUY FC 24 COINS

compras monedas fc 24

Volunteering at Soi Dog

Find a Job

Jooble jobs

Free ibomma Movies