[新聞] TikTok看到大量入侵烏克蘭聲稱是誤導假影片

看板Gossiping作者Ruo5566 (若)時間2年前 (2022/02/27 10:13)推噓1(2推 1噓 10→)

留言13則, 9人參與討論串1/1

https://reurl.cc/X437Rj TikTok看到大量入侵烏克蘭聲稱是誤導假影片 February 25, 2022 10:59 PM ET Bobby Allyn https://i.imgur.com/UK9N2o0.jpg

The invasion of Ukraine has seen a surge of videos flooding TikTok, many of the most popular ones containing false or misleading material. Kiichiro Sato/AP Russia's invasion of Ukraine has unleashed a flood of misleading and false material on TikTok. The popular app used by more than 1 billion people has been amplifying videos portraying old conflicts, scenes from movies and even video game battles as if showing on-the-ground live footage. In times of crisis, social media platforms are always struggling to stay head of misinformation and make round-the-clock calls on when a viral post should be removed. But the flurry of conflict-themed footage now on TikTok has overwhelmed the platform in new ways, sending countless fake or videos framed as if depicting the war in Ukraine to millions of viewers. "Though it's crucial that the public remain informed of such high-stakes situations, it seems that the platform's design is incompatible with the needs of the current moment," wrote Abbie Richards of the liberal watchdog group Media Matters. As of Friday evening, videos with the hashtag #RussianInvasion have received 32 million views and videos with the hashtag #RussiaUkraine have racked up 132 million views. "This is the first time TikTok has really been central in a conflict situation of this scale," said Sam Gregory, the program director of Witness, a nonprofit focused on the ethical use of video in humanitarian crises. "And the volume of misleading videos does seem new to me. Some people are doing it because they want attention, some people want to monetize it, others are doing it potentially as misinformation and disinformation," he said. Some users are exploiting features that help videos on TikTok go viral, including reusing an audio clip with new footage. Audio of gunfire uploaded from before the war started was used in more than 1,700 videos before it was removed, often featuring shaky camera footage to give the impression that it was capturing a conflict, according to Media Matters. The group also found that a video featuring audio from a 2020 explosion in Beirut was watched more than 6 million times in just 12 hours. TikTok's community guidelines say it bans misinformation "that causes harm to individuals," such as videos that incite hate or prejudice. But footage misrepresenting scenes of war does not appear to explicitly violate the company's content policies. TikTok did not return a request for comment. In the past two days, an NPR reporter was served up an unlabeled video showing a movie depiction of war that was viewed nearly 50 million times and prompted a conversation in the comments about whether it was Ukraine. An old Albanian training exercise purporting to show Ukraine that was seen almost 15 million times. And a 2014 video watched on the platform about 5 million times claiming to show Ukrainian and Russian soldiers "face to face." TikTok's design creates something of a paradox: If you watch a video repeatedly to try to decipher whether it's authentic or return to one after conducting research, "you're telling the algorithm you want more of this," Gregory said. Researchers like Gregory say TikTok can do more to give users tools to quickly figure out if a video is fake: the ability to do instant reverse image searches to see whether the video has circulated in the past and databases where users can go to see if popular videos have already been debunked. Often if a video is fraudulent, TikTok commenters will point it out and the comment will rise to the top of the video's discussion section, but waiting for a TikTok user to figure out if a video is fake is often too little, too late, researchers say. Gregory said TikTok has the potential to make humanitarian crises and wars more vivid and tangible to a massive audience who may not have otherwise engaged at all, but time spent watching fake videos of war does little to add to a person's understanding of a conflict. "We shouldn't reject instantly that ephemeral moments in peoples' lives is bad. There can be ways TikTok can help people engage and have dialogue with people on the front lines," Gregory said. "But the challenge is finding those moments within all the manipulation." -----估狗翻譯----- 入侵烏克蘭後，TikTok 湧現了大量視頻，其中許多最受歡迎的視頻包含虛假或誤導性材料。佐藤喜一郎/美聯社俄羅斯入侵烏克蘭在 TikTok 上釋放了大量誤導性和虛假材料。超過 10 億人使用的流行應用程序一直在放大描繪舊衝突、電影場景甚至電子遊戲戰鬥的視頻，就好像在顯示現場實況鏡頭一樣。在危機時期，社交媒體平台一直在努力保持錯誤信息的領先地位，並全天候呼籲何時刪除病毒式帖子。但是，現在 TikTok 上的一系列以沖突為主題的視頻以新的方式淹沒了該平台，向數百萬觀眾發送了無數假冒或框架化的視頻，彷彿描繪了烏克蘭的戰爭。自由派監督組織 Media Matters 的 Abbie Richards寫道： “儘管讓公眾隨時了解這種高風險情況至關重要，但該平台的設計似乎與當前的需求不相容。” “這是 TikTok 第一次真正在這種規模的衝突局勢中發揮核心作用，”Witness 的項目主管 Sam Gregory 說，該組織專注於人道主義危機中視頻的合乎道德使用。 “而且誤導性視頻的數量對我來說確實很新鮮。有些人這樣做是因為他們想要關注，有些人想要將其貨幣化，其他人則可能將其作為錯誤信息和虛假信息，”他說。一些用戶正在利用幫助 TikTok 上的視頻傳播的功能，包括重複使用帶有新鏡頭的音頻剪輯。據媒體報導，在戰爭開始前上傳的槍聲音頻在被刪除之前被用於 1,700 多個視頻中，通常帶有搖晃的攝像機鏡頭，給人的印像是它正在捕捉一場衝突。該組織還發現，一段包含 2020 年貝魯特爆炸音頻的視頻在短短 12 小時內被觀看了超過 600 萬次。 TikTok 的社區指南稱，它禁止“對個人造成傷害”的錯誤信息，例如煽動仇恨或偏見的視頻。但歪曲戰爭場面的鏡頭似乎並未明確違反公司的內容政策。在過去的兩天裡，一位 NPR 記者收到了一段未標記的視頻，該視頻顯示了一部描述戰爭的電影，該電影被觀看了近 5000 萬次，並在評論中引發了關於是否是烏克蘭的對話。一個古老的阿爾巴尼亞訓練演習，旨在向烏克蘭展示近 1500 萬次觀看。2014 年，該平台上的一段視頻被觀看了大約 500 萬次，聲稱顯示烏克蘭和俄羅斯士兵“面對面”。 TikTok 的設計創造了一個悖論：如果你反復觀看一段視頻以試圖破譯它是真實的還是在進行研究後返回一個視頻，“你就是在告訴算法你想要更多這樣的東西，”格雷戈里說。像 Gregory 這樣的研究人員表示，TikTok 可以做更多的事情來為用戶提供快速判斷視頻是否為假的工具：能夠進行即時反向圖像搜索以查看視頻是否在過去傳播過，以及用戶可以訪問的數據庫以查看是否流行視頻已經被揭穿。通常，如果視頻是欺詐性的，TikTok 評論者會指出這一點，評論將上升到視頻討論部分的頂部，但等待 TikTok 用戶弄清楚視頻是否是假的往往太少，太遲了，研究人員說。 Gregory 說，TikTok 有可能使人道主義危機和戰爭對可能根本沒有參與的大量觀眾更加生動和切實，但花時間觀看假戰爭視頻並不能增加人們對沖突的理解。 “我們不應該立即否認人們生活中的短暫時刻是不好的。TikTok可以通過多種方式幫助人們與前線的人們互動並進行對話，”格雷戈里說。“但挑戰在於在所有操縱中找到那些時刻。” -- ※ 發信站: 批踢踢實業坊(ptt.cc), 來自: 111.252.65.201 (臺灣) ※ 文章網址: https://www.ptt.cc/bbs/Gossiping/M.1645927995.A.6C2.html

→

Aotearoa

02/27 10:13, 2年前 , 1^F

02/27 10:13, 1^F

→

YLTYY

02/27 10:14, 2年前 , 2^F

02/27 10:14, 2^F

推

ggbi4zzz

02/27 10:14, 2年前 , 3^F

02/27 10:14, 3^F