Feb 05, 2020 11:02:00

Twitter announces that it will label `` edited movies and photos '' to prevent people from being fooled

It is not unusual to see photos and movies edited by users on SNS, and many people will be looking forward to seeing interesting Photoshop images and movies. However, Twitter acknowledges that edited content may hurt people or pose a threat to certain groups, and a new statement, `` Labeling edited movies and photos to deceive people. Not be done. '

New measures for synthesized or manipulated media

https://blog.twitter.com/ja_jp/topics/company/2019/Building-rules-in-public-ja.html

From late October 2019, Twitter has collected more than 6,500 opinions using a short survey and hashtags on the platform, how to deal with edited movies and photos . In addition, after exchanging opinions with various groups such as civil society groups and academic researchers, we found that 'Through the overall threat of misleading and falsified content has been recognized, Twitter will take measures against them. It ’s now clear that you ’re asking for it. ”

According to the survey, more than 70% of Twitter users answered that 'it is unacceptable to take' no measures 'for content that has been tampered with and misleading,' He said that it was acceptable to attach a 'warning label'. On the other hand, many people were more reluctant to make edited content completely unpostable, and some commented that there were issues such as restrictions on expression and censorship.

Therefore, Twitter has decided to label the edited movie or photo so that viewers can immediately determine that this is being edited.

We want you to be able to see and share information on Twitter without worry. However, it can be difficult to determine what is real in the videos and photos on the timeline. To counter such erroneous information, we will establish new rules and update products accordingly. pic.twitter.com/61DgU7nyXA

— Twitter Japan (@TwitterJP) February 4, 2020

At the time of writing, to determine whether the content you saw for the first time has been edited, it is necessary to comprehensively determine factors such as context and sender.

However, with new measures ...

Edited content will be labeled 'Media manipulated'.

Tap the label attached to the content ...

You can see external information sources and check how the content has been tampered with.

Twitter also has a new rule for 'synthesized or manipulated media.' The '

Twitter Rules ' page states that 'It is prohibited to share synthetic or manipulated media that is likely to cause any damage in an attempt to deceive you. We may label tweets that contain media that has been manipulated or manipulated to help users understand the authenticity and background information of the tweet. '

Twitter uses the following three criteria for labeling.

◆ 1: Content synthesized or manipulated
`` Are significant edits made that fundamentally change the composition, flow, timing, and framework? '' `` Add or remove visual or audio information (new video frames, audio dubbing, subtitle changes, etc.) Twitter determines whether the content has been significantly tampered with, based on factors such as whether the content has been forged or imitated.

◆ 2: Content is shared in such a way that it distracts others
Twitter says that the context and background information in which the content is shared are also used as judgment factors, and it is said that 'text of the tweet attached to the content or text included in the content' 'metadata accompanying the content' 'content The company will examine the background provided with the content, such as the profile information of the person sharing the content, and the website linked to the profile of the person sharing the content and the tweet sharing the content.

◆ 3: Can content threaten public safety and cause serious harm?
Threats to the physical security of individuals and groups, Risks of collective violence and riots, Threats to privacy of individuals and groups, free expression and participation in public meetings, such as stalking, unwanted threats, and specific targets Content that threatens public safety and causes serious harm, such as metaphors, swearing, or material that silences or oppress or intimidate voters. It is targeted.

The chart for content labeling or removal criteria is as follows:

Twitter will begin labeling tweets on March 5, 2020, saying, `` This is a challenge for us, and it is possible that mistakes may occur in implementing them, but we have a patient relationship I would be grateful if you could. '