プラットフォームサービスに関する研究会(第49回)開催案内
一般職事務系(高卒・行政評価)の採用情報を更新しました。
「統計データ・グラフフェア」の開催
消費者保護ルールの在り方に関する検討会(第50回)
行政評価局総務課採用担当 選考採用職員採用情報
[B] フォルカー・シュレンドルフ監督『シャトーブリアンからの手紙』(2011年)
【焦点】インボイス反対署名36万提出 経済悪化を招く消費増税 稀代の悪法中止・延期を=橋詰雅博
EFF's Comment to the Meta Oversight Board on Polish Anti-Trans Facebook Post
EFF recently submitted comments in response to the Meta Oversight Board’s request for input on a Facebook post in Polish from April 2023 that targeted trans people. The Oversight Board was created by Meta in 2020 as an appellate body and has 22 members from around the world who review contested content moderation decisions made by the platform.
Our comments address how Facebook’s automated systems failed to prioritize content for human review. From our observations—and the research of many within the digital rights community—this is a common deficiency made worse during the pandemic, when Meta decreased the number of workers moderating content on its platforms. In this instance, the content was eventually sent for human review and was still assessed to be non-violating and therefore not escalated further. Facebook kept the content online despite 11 different users reporting the content 12 times and only removed the content once the Oversight Board decided to take the case for review.
As EFF has demonstrated, Meta has at times over-removed legal LGBTQ+ related content whilst simultaneously keeping content online that depicts hate speech toward the LGBTQ+ community. This is often because the content—as in this specific case—is not an explicit depiction of such hate speech, but rather a message that is embedded in a wider context that automated content moderation tools and inadequately trained human moderators are simply not equipped to consider. These tools do not have the ability to recognize nuance or the context of statements, and human reviewers are not provided the training to remove content that depicts hate speech beyond a basic slur.
This incident serves as part of the growing body of evidence that Facebook’s systems are inadequate in detecting seriously harmful content, particularly that which targets marginalized and vulnerable communities. Our submission looks at the various reasons for these shortcomings and makes the case that Facebook should have removed the content—and should keep it offline.
Read the full submission in the PDF below.
Surge in police use of facial recognition sparks concerns over wrongful targeting
Chris Jones, director of Statewatch, said: “There is a significant possibility for innocent people to be wrongly identified and potentially even charged and convicted through the use of this technology.”
Mr Jones warned that use of RFR could perpetuate discriminatory profiling: “We should also be asking who those people are most likely to be.
“The Casey Review confirmed that the Metropolitan Police are institutionally racist, misogynistic and homophobic, while Avon and Somerset’s own chief constable has admitted her force is institutionally racist.”
Full story here.
「こんな理由で解雇されてはたまらない」〜若き西武バス運転手の裁判を傍聴して
JVN: Advantech製EKI-1524-CE シリーズなどにおける複数のクロスサイトスクリプティングの脆弱性
Googleマップに壊れた橋を通る経路を案内され事故死したとして、遺族がGoogleらを訴える
すべて読む | ITセクション | Google | 法廷 | IT | 交通 |
関連ストーリー:
岩手県の奇妙な石碑、Googleマップ登録をきっかけに話題に 2023年07月05日
2023年3月21日にストリートビューアプリがサポート終了する説出回る 2022年11月07日
路線バスの乗り方が分からなかった検索しろは正しいか? 2022年09月27日
Googleのロシア法人が破産、当局による銀行口座差し押さえで 2022年05月24日
お知らせ:CyberNewsFlash「Apple製品のアップデートについて(2023年9月 第2号)」(更新)
景気動向指数(令和5年7月分速報からの改訂状況)
西友が新システムへ切替で店頭在庫の品薄・欠品が発生
関連ストーリー:
IKEA、欧州の店舗で在庫管理に用いるドローンが 100 機に到達 2023年03月21日
森永製菓に不正アクセス、一部商品の製造が影響を受ける 2022年03月19日
日経、米ウォルマートが西友を売却すると報道 2018年07月14日
大手小売り、相次いで値下げを発表 2017年08月31日