Connect with us

SOCIAL

Facebook Outlines New, Machine Learning Process to Improve the Accuracy of Community Standards Enforcement

Published

on

Facebook is always working to improve its detection and enforcement efforts in order to remove content that breaks its rules, and keep users safe from abuse, misinformation, scams, etc. 

And it’s systems have significantly improved in this respect – as explained by Facebook:

Online services have made great strides in leveraging machine-learned models to fight abuse at scale. For example, 99.5% of takedowns on fake Facebook accounts are proactively detected before users report them.”

But there are still significant limitations in its processes, mostly due to the finite capacity for human reviewers to assess and pass judgment on such instances. Machine learning tools can identify a growing number of issues, but human input is still required to confirm whether many of those identified cases are correct, because computer systems often miss the complex nuance of language.

But now, Facebook has a new system to assist in this respect:

CLARA (Confidence of Labels and Raters), is a system built and deployed at Facebook to estimate the uncertainty in human-generated decisions. […] CLARA is used at Facebook to obtain more accurate decisions overall while reducing operational resource use.” 

The system essentially augments human decision making by adding a machine learning layer on top of that which assesses each individual raters’ capacity to make the right call on content, based on their past accuracy.

Facebook CLARA system

The CLARA element is in the ‘Realtime Prediction Service’ sector of this flow chart, which assesses the result of each incident and crosschecks the human ruling against what the machine model would have predicted, while also referencing both against each reviewers’ past results for each type of report.   

See also  HQ Trivia shuts down after acquisition falls through

That system, which has now been deployed at Facebook, has resulted in a significant improvement in efficiency, ensuring more accurate results in enforcement.

Compared to a random sampling baseline, CLARA provides a better trade-off curve, enabling an efficient usage of labeling resources. In a production deployment, we found that CLARA can save up to 20% of total reviews compared to majority vote.”

Which is important right now, because Facebook has been forced to reduce its human moderation capacity due to COVID-19 lockdowns in different regions. By improving its systems for accurately detecting violations, through automated means, Facebook is then able to concentrate its resources on the key areas of concern, maximizing the manpower that it has available.

Of course, there are still issues with Facebook’s systems. Just this week, reports emerged that Facebook is looking at a new way to use share velocity signals in order to better guide human moderation efforts, and stop misinformation, in particular, from reaching massive audiences on the platform. That comes after a recent COVID-19 conspiracy video was viewed some 20 million times on Facebook, despite violating the platform’s rules, before, finally, Facebook’s moderators moved to take it down.

Improved human moderation wouldn’t have helped in this respect, so there are still other areas of concern for Facebook to address, but the smarter it can be about utilizing the resources it has available, the more Facebook can focus its efforts onto the key areas of concern, in order to detect and remove violating content before it can reach big audiences.

You can read more about Facebook’s new CLARA reinforcement process here

Socialmediatoday.com

Continue Reading
Advertisement

SOCIAL

Meta’s Developing and ‘Ethical Framework’ for the Use of Virtual Influencers

Published

on

Meta's Developing and 'Ethical Framework' for the Use of Virtual Influencers


With the rise of digital avatars, and indeed, fully digital characters that have evolved into genuine social media influencers in their own right, online platforms now have an obligation to establish clear markers as to what’s real and what’s not, and how such creations can be used in their apps.

The coming metaverse shift will further complicate this, with the rise of virtual depictions blurring the lines of what will be allowed, in terms of representation. But with many virtual influencers already operating, Meta is now working to establish ethical boundaries on their application.

As explained by Meta:

From synthesized versions of real people to wholly invented “virtual influencers” (VIs), synthetic media is a rising phenomenon. Meta platforms are home to more than 200 VIs, with 30 verified VI accounts hosted on Instagram. These VIs boast huge follower counts, collaborate with some of the world’s biggest brands, fundraise for organizations like the WHO, and champion social causes like Black Lives Matter.”

Some of the more well-known examples on this front are Shudu, who has more than 200k followers on Instagram, and Lil’ Miquela, who has an audience of over 3 million in the app.

At first glance, you wouldn’t necessarily realize that this is not an actual person, which makes such characters a great vehicle for brand and product promotions, as they can be utilized 24/7, and can be placed into any environment. But that also leads to concerns about body image perception, deepfakes, and other forms of misuse through false or unclear representation.

See also  Facebook Looks to Boost Watch Viewership with Exclusive Deal for Indian Cricket Tour

Deepfakes, in particular, may be problematic, with Meta citing this campaign, with English football star David Beckham, as an example of how new technologies are evolving to expand the use of language, as one element, for varying purpose.

The well-known ‘DeepTomCruise’ account on TikTok is another example of just how far these technologies have come, and it’s not hard to imagine a scenario where they could be used to, say, show a politician saying or doing something that he or she actually didn’t, which could have significant real world impacts.

Which is why Meta is working with developers and experts to establish clearer boundaries on such use – because while there is potential for harm, there are also beneficial uses for such depictions.

Imagine personalized video messages that address individual followers by name. Or celebrity brand ambassadors appearing as salespeople at local car dealerships. A famous athlete would make a great tutor for a kid who loves sports but hates algebra.

Such use cases will increasingly become the norm as VR and AR technologies are developed, with these platforms placing digital characters front and center, and establishing new norms for digital connection.

It would be better to know what’s real and what’s not, and as such, Meta needs clear regulations to remove dishonest depictions, and enforce transparency over VI use.

But then again, much of what you see on Instagram these days is not real, with filters and editing tools altering people’s appearance well beyond what’s normal, or realistic. That can also have damaging consequences, and while Meta’s looking to implement rules on VI use, there’s arguably a case for similar transparency in editing tools applied to posted videos and images as well.

See also  TikTok Adds Sticker Pinning in Video Clips

That’s a more complex element, particularly as such tools also enable people to feel more comfortable in posting, which no doubt increases their in-app activity. Would Meta be willing to put more focus on this element if it could risk impacting user engagement? The data on the impact of Instagram on people’s mental health are pretty clear, with comparison being a key concern.

Should that also come under the same umbrella of increased digital transparency?

It’s seemingly not included in the initial framework as yet, but at some stage, this is another element that should be examined, especially given the harmful effects that social media usage can have on young women.

But however you look at it, this is no doubt a rising element of concern, and it’s important for Meta to build guardrails and rules around the use of virtual influencers in their apps.

You can read more about Meta’s approach to virtual influencers here.





Source link

Continue Reading

SOCIAL

Meta Publishes New Guide to the Various Security and Control Options in its Apps

Published

on

Meta Publishes New Guide to the Various Security and Control Options in its Apps


Meta has published a new set of safety tips for journalists to help them protect themselves in the evolving online connection space, which, for the most part, also apply to all users more broadly, providing a comprehensive overview of the various tools and processes that it has in place to help people avoid unwanted attention online.

The 32-page guide is available in 21 different languages, and provides detailed overviews of Meta’s systems and profile options for protection and security, with specific sections covering Facebook, Instagram and WhatsApp.

The guide begins with the basics, including password protections and enabling two-factor authentication.

It also outlines tips for Page managers in securing their business profiles, while there are also notes on what to do if you’ve been hacked, advice for protection on Messenger and guidance on bullying and harassment.

Meta security guide

For Instagram, there are also general security tips, along with notes on its comment moderation tools.

Meta security guide

While for WhatsApp, there are explainers on how to delete messages, how to remove messages from group chats, and details on platform-specific data options.

Meta security guide

There are also links to various additional resource guides and tools for more context, providing in-depth breakdowns of when and how to action the various options.

It’s a handy guide, and while there are some journalist-specific elements included, most of the tips do apply to any user, so it could well be a valuable resource for anyone looking to get a better handle on your various privacy tools and options.

Definitely worth knowing either way – you can download the full guide here.

See also  If you love voice messaging, you’ll love Squad



Source link

Continue Reading

SOCIAL

Twitter bans account linked to Iran leader over video threatening Trump

Published

on

Twitter bans account linked to Iran leader over video threatening Trump


Iran’s supreme leader Ayatollah Ali Khamenei meets with relatives of slain commander Qasem Soleimani ahead of the second anniverary of his death in a US drone strike in Iraq – Copyright POOL/AFP/File Tom Brenner

Twitter said Saturday it had permanently suspended an account linked to Iran’s supreme leader that posted a video calling for revenge for a top general’s assassination against former US president Donald Trump.

“The account referenced has been permanently suspended for violating our ban evasion policy,” a Twitter spokesperson told AFP.

The account, @KhameneiSite, this week posted an animated video showing an unmanned aircraft targeting Trump, who ordered a drone strike in Baghdad two years ago that killed top Iranian commander General Qassem Soleimani.

Supreme leader Ayatollah Ali Khamenei’s main accounts in various languages remain active. Last year, another similar account was suspended by Twitter over a post also appearing to reference revenge against Trump.

The recent video, titled “Revenge is Definite”, was also posted on Khamenei’s official website.

According to Twitter, the company’s top priority is keeping people safe and protecting the health of the conversation on the platform.

The social media giant says it has clear policies around abusive behavior and will take action when violations are identified.

As head of the Quds Force, the foreign operations arm of Iran’s Revolutionary Guards, Soleimani was the architect of its strategy in the Middle East.

He and his Iraqi lieutenant were killed by a US drone strike outside Baghdad airport on January 3, 2020.

Khamenei has repeatedly promised to avenge his death.

On January 3, the second anniversary of the strike, the supreme leader and ultraconservative President Ebrahim Raisi once again threatened the US with revenge.

See also  Pinterest Launches New Holiday Marketing Hub to Assist with Your Campaign Planning

Trump’s supporters regularly denounce the banning of the Republican billionaire from Twitter, underscoring that accounts of several leaders considered authoritarian by the United States are allowed to post on the platform.



Source link

Continue Reading

DON'T MISS ANY IMPORTANT NEWS!
Subscribe To our Newsletter
We promise not to spam you. Unsubscribe at any time.
Invalid email address

Trending