- 谁在运行
- Department for Transport
- 它会影响人们吗？
- 有可能。它可能影响人们看到、获推荐、被标记或被优先处理的内容。
- 可能的权利影响: 间接公共影响
我们了解多少？
来源是信息所依据的公开网页或文件。标记说明记录有多完整，不代表系统是否良好或安全。缺少信息不等于缺少保护措施。
来源与延伸阅读
- DfT: EV Chargepoint Grant Photos Fraud Tool / GOV.UK Algorithmic Transparency Records / 30 July 2026
- GOV.UK Algorithmic Transparency Records - official inventory / GOV.UK Algorithmic Transparency Records
可选详情完整记录与详情
界面和说明指南已经翻译。记录名称与描述、引文、详细编辑说明和链接的公共文件保留发布时的语言，避免悄然改变证据。
它做什么
A deep-learning image tool that flags duplicate or web-sourced installation photos submitted to electric-vehicle chargepoint grant schemes.
背景说明
Two grant claims may contain almost the same installation photo. The model brings that pair to an administrator's attention instead of treating similarity as proof of fraud.
Human decides every flagged image is checked by a grant administrator
待核事实
- 01
The system uses image recognition and Google Vision API.
- 02
It looks for duplicate installations and images taken from the web.
- 03
It cannot approve or reject a grant on its own.
如果它影响到个人
A flag can lead to closer checking of evidence in an EV chargepoint grant application.
Grant administrators inspect every flagged image and make the final decision. The record explicitly says the tool has no decision-making capacity.
The existing grant appeal mechanisms still apply because administrators, not the model, decide the outcome.
Photos submitted as evidence to EV chargepoint grant schemes and image-similarity signals used to compare them.
来源与延伸阅读
来源是支撑某项说法的公共文件。请直接打开，核实其内容和发布者。
通俗文本已对照来源核查。旧记录未保留模型或提示日志，因此它只是说明，不是证据。
待核事实
- 谁在运行
- Department for Transport
- 生活领域
- 公共部门
- 证据级别
- 来源支持
- 注释
- 2
- 更新于
- 48天前
分数安排工作，不评价人。
涉及权利的敏感用途、缺少来源、未回答的问责问题和陈旧证据会提高审查优先级；有力的官方文件会降低它。该分数不是对安全、公平或合法性的裁决。
空白字段表示所附公共来源没有给出答案。覆盖范围并不均衡，公开文件不能证明实际表现，证书也有明确适用边界。
以前的版本
- Imported15天前较充分Imported from GOV.UK Algorithmic Transparency Records.
- Public register record48天前较充分Added from GOV.UK Algorithmic Transparency Records.
帮助核查这条记录。
添加公共来源、质疑某项说法，或指出缺少联系人工的途径。
提交的姓名、邮箱、说明、来源或更正仅用于核实记录和防止滥用，不会自动公开。