User awareness and defenses against sockpuppet friend invitations in Facebook
User awareness and defenses against sockpuppet friend invitations in Facebook
复制标题
用户对 Facebook 中马甲好友邀请的认知和防御
DOI：
10.1145/3477314.3507012
复制
发表时间：
2022
期刊：
影响因子：
--
通讯作者：
Carbunar, Bogdan
中科院分区：
文献类型：
--
作者：
Talukder, Sajedul;Hernandez, Nestor;Azimpourkivi, Mozhgan;Carbunar, Bogdan
Facebook中的朋友关系已被用来访问私人和敏感的用户数据，在受害者的时间表和新闻源上发布虚假或辱骂性信息，以及诈骗和影响受害者的看法。在对35名使用Facebook的参与者进行的定性研究中，我们发现他们经常确认从sockpuppet帐户收到的邀请。一些参与者声称以前接触过这种邀请，并表现出更高的认识。然而，在我们的研究中，几个这样的参与者也确认了至少一个sockpuppet邀请。为了采取措施解决这个问题，在本文中，我们引入了一个新的界面来通知，教育和推动用户在做出决定之前检查他们的Facebook好友邀请，并避免确认从可疑帐户收到的邀请。我们提出了一个分类器来检测不受欢迎的朋友邀请，并使用它来优先考虑一个必看的未决朋友列表。在对145名参与者的研究中，我们发现，与对照的类似Facebook的界面相比，我们的解决方案将来自sockpuppet帐户的确认好友邀请减少了42.6%，将此类帐户的盲确认减少了22.6%，并将此类帐户的检查配置文件增加了54.67个百分点。我们表明，当训练只有少数用户的决定，从这些研究中，我们的分类器预测用户的决定未决的朋友与96.52%的F1测量。
Friend relations in Facebook have been used to access private and sensitive user data, post false or abusive information on the timelines and news feeds of victims, and scam and influence the perceptions of victims. In a qualitative study with 35 participants who use Facebook we found that they often confirm invitations received from sockpuppet accounts. Some participants claimed prior exposure to such invitations and exhibited increased awareness. However, in our study several such participants have also confirmed at least one sockpuppet invitation. To take steps toward addressing this problem, in this paper we introduce a new interface to inform, educate and nudge users to inspect their Facebook friend invitations before making a decision, and avoid confirming invitations received from suspicious accounts. We propose a classifier to detect undesirable friend invitations and use it to prioritize a must-see pending friend list. In studies with 145 participants, we found that when compared to a control, Facebook-like interface, our solution reduced confirmed friend invitations from sockpuppet accounts by 42.6 percentage points, reduced blind confirmations of such accounts by 22.6 percentage points, and increased the inspected profiles of such accounts by 54.67 percentage points. We show that when trained on only a few user decisions from these studies, our classifier predicted the user decisions for pending friends with an F1-measure of 96.52%.
登录
查看更多内容
DOI：
--
发表时间：
1977
期刊：
The Hastings Center report
影响因子：
--
作者：
通讯作者：
--
去 SPB 还是不去 SPB？
影响因子：
6.1
作者：
Z. Vakhitova;R. Mawby;C. Alston;Callum A. Stephens
通讯作者：
Callum A. Stephens
社交媒体的隐私推动：Facebook 的一项探索性研究
DOI：
--
发表时间：
2013
期刊：
The Web Conference
影响因子：
--
作者：
Yang Wang;P. Leon;Kevin Scott;Xiaoxuan Chen;A. Acquisti;L. Cranor
通讯作者：
L. Cranor
检查垃圾邮件的需求：谁点击？
DOI：
10.1145/3173574.3173786
发表时间：
2018
期刊：
Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems
影响因子：
--
作者：
Elissa M. Redmiles;Neha Chachra;Brian Waismeyer
通讯作者：
Brian Waismeyer