All legal documents

AI Addendum

TimePlate — SlashHub Limited

TimePlate AI Addendum

Last updated: 2026-10-12 · Version 1.0.0

What this is. This is the contractual addendum for AI features in TimePlate. It attaches to, and forms part of, the Master Services Agreement ("MSA") and the Terms of Service. It is different from the user-facing AI Disclosure (a transparency notice rendered at /ai): the disclosure says what the AI is; this Addendum says who owns what, what we may train on, and who bears the risk. Precedence. On AI matters, this Addendum prevails over the MSA and the Terms of Service; in all other respects the MSA governs. Language. Drafted in English with a 繁體中文 translation. The English version governs unless a signed contract states otherwise.

1. Definitions

  • "AI Features" — artificial-intelligence and machine-learning features in the Services, including the SlashAI agent, natural-language assistance, shift/schedule suggestions, demand forecasting, labour-cost optimisation, payroll/attendance anomaly detection, smart scheduling and any feature we describe as "AI".
  • "Input" — data, text, files or instructions that the Customer or its Authorised Users submit to an AI Feature.
  • "Output" — the response or result an AI Feature generates.
  • "Model Provider" — a third party whose model the AI Features call.
  • "Training" — using data to develop, fine-tune, calibrate or improve a model, algorithm or recommendation engine.

2. What the AI Features are — and are not

2.1 The AI Features are assistive tools. They produce drafts, suggestions, predictions and computations. They are not sentient, are not professional advisers, and are not a substitute for human judgement. Output may be inaccurate, incomplete, biased, out of date or fabricated ("hallucination").

2.2 The Customer must ensure that a competent human reviews and approves every Output before it is relied on, published to employees, used in payroll or filing, or used in any decision that materially affects an individual. The Customer must not use the AI Features to make an employment decision (including hiring, scheduling that affects pay or rest, discipline, demotion or termination) without meaningful human review.

3. Ownership of Output

3.1 Input. The Customer (or its licensors) retains all rights in Input. The Customer grants SlashHub a licence to process Input solely to provide the AI Features and as permitted by the DPA.

3.2 Output — ownership. As between SlashHub and the Customer, the Customer owns the Output to the extent it can be owned. SlashHub assigns to the Customer all rights it may have in Output that is generated for the Customer and does not claim ownership of it. This assignment covers only Output generated for the Customer, transfers no SlashHub IP (or IP of its licensors), and excludes any third-party or Model-Provider content embedded in the Output, which remains subject to that party's terms.

3.3 No warrant of protectability or uniqueness. Because of the nature of generative AI, SlashHub does not warrant that Output is protectable by copyright or other IP, is unique, or does not resemble output generated for others or existing third-party material. The Customer is responsible for assessing whether Output is suitable to use, protect or register.

3.4 Licence back to the Customer. Subject to payment and compliance, SlashHub grants the Customer a worldwide, non-exclusive, royalty-free, revocable licence to use, reproduce, modify and distribute Output for the Customer's internal business purposes. This licence does not permit using Output to train a competing AI model, or representing AI Output as human-authored where that would be misleading or is required to be disclosed.

4. Model-training permissions

4.1 We do not train our own models on Customer Data. SlashHub does not use Customer Data (including Personal Data of employees) to train, fine-tune or improve any AI model of its own, and does not sell Customer Data. This reflects the DPA.

4.2 Model Providers. SlashHub contracts with Model Providers on terms that (a) prohibit them from using Customer Data to train their general models, and (b) impose confidentiality and security obligations no less protective than the DPA. Where a Model Provider cannot meet these terms for a given feature, SlashHub will not route Personal Data to that feature.

4.3 Anonymised improvement (opt-in). SlashHub may use aggregated, anonymised data that cannot reasonably be linked to any individual to operate, secure and improve the Services. If the Customer wishes to opt out of any optional anonymised-improvement programme, it may do so in account settings or by writing to ai@slashhub.hk.

4.4 Customer's own training data. The Customer shall not use the AI Features to train, fine-tune or distil a competing model, and shall not submit Input it does not have the right to submit.

5. Human oversight and transparency

5.1 Every irreversible AI-driven action (for example creating a record, sending a message, or exporting payroll) requires the Customer's or an Authorised User's confirmation, or is logged and reversible.

5.2 AI involvement is labelled in the interface, and AI tool calls are recorded in the account activity log. Where the Customer publishes or shares Output, the Customer is responsible for disclosing its AI origin where the law requires.

6. Accuracy, verification and liability

6.1 No accuracy warranty. SlashHub provides no warranty as to the accuracy, completeness, currency or fitness for purpose of any Output. The Customer must independently verify every Output before relying on it, particularly payroll, MPF, statutory-holiday, leave, overtime, tax and IR56 figures.

6.2 Liability. SlashHub's liability for the AI Features is subject to the exclusions and cap in the MSA clause 12 (and MSA clause 12.3 carve-outs). SlashHub is not liable for any decision a person makes in reliance on Output, or for any loss arising from Output that was not verified as clause 6.1 requires.

6.3 Customer indemnity. The Customer shall indemnify SlashHub against claims arising from the Customer's use of Output (including an automated decision or an employment decision made without the review required by clauses 2.2 and 5.1).

7. EU AI Act and regulated uses

7.1 SlashHub classifies the TimePlate AI Features as limited-risk under the EU AI Act, and relies on transparency and human oversight (this Addendum and the AI Disclosure). They are not intended to be prohibited AI practices, and SlashHub does not intend them to be used for biometric categorisation or emotion recognition.

7.2 Employment context. The Customer acknowledges that AI used in an employment context can be high-risk under the EU AI Act (Annex III). Where the Customer is established in, or targets individuals in, the EEA, the Customer is responsible for any additional obligations that apply, including human oversight, data governance, registration and any fundamental-rights impact assessment, and must configure the AI Features accordingly (assistive, human-reviewed). SlashHub offers the AI Features as assistive tools only and does not warrant compliance with the Customer's sector-specific obligations. Customers with regulated or high-risk use cases should contact ai@slashhub.hk.

8. Beta AI features

AI Features labelled beta/preview/experimental are provided "as is", may change or be withdrawn, carry no service-level commitment, and should not be used with real payroll or other consequential data without the Customer's own testing.

9. Term, changes and contact

9.1 This Addendum runs with the MSA and is governed by its governing-law and dispute-resolution clauses (Hong Kong law; HKIAC arbitration, seat Hong Kong).

9.2 SlashHub may update this Addendum as described in MSA clause 19, on 30 days' notice for material changes.

9.3 AI enquiries and reports: ai@slashhub.hk.

TimePlate AI 附件

最後更新:2026-10-11 · 版本 1.0.0(草擬本) · 狀態:範本 — 待香港法律審閱

請先閱讀。 本文件為借助 AI 起草的範本,並非法律意見,使用前必須由合資格香港律師審閱及簽署。 本文件是什麼。 本文件是 TimePlate 人工智能功能的合約附件,附於並構成主服務協議(「MSA」)及服務條款的一部分。它與面向使用者的 AI 披露聲明(載於 /ai 的透明度聲明)不同:披露聲明說明 AI 是什麼;本附件說明 誰擁有什麼、我們可就什麼進行訓練,以及由誰承擔風險。 優先次序。 就 AI 事宜,本附件優先於 MSA 及服務條款;在其他方面則以 MSA 為準。 語言。 本文件以英文草擬並附繁體中文譯本。除經簽署的合約另有訂明外,以英文版本為準。

1. 定義

  • 「AI 功能」 — 服務中的人工智能及機器學習功能,包括 SlashAI 代理、自然語言協助、更表/排程建議、需求預測、勞動力成本優化、薪酬/出勤異常偵測、智能排更,以及我們描述為「AI」的任何功能。
  • 「輸入」 — 客戶或其授權使用者向 AI 功能提交的資料、文字、檔案或指示。
  • 「輸出」 — AI 功能產生的回應或結果。
  • 「模型供應商」 — AI 功能所呼叫的第三方模型之供應商。
  • 「訓練」 — 使用資料開發、微調、校準或改進模型、演算法或推薦引擎。

2. AI 功能是什麼 — 及不是什麼

2.1 AI 功能為輔助工具,產生草稿、建議、預測及計算。它們並非有感知,並非專業顧問,亦非人手判斷的替代品。輸出可能不準確、不完整、有偏見、過時或屬虛構(「幻覺」)。

2.2 客戶必須確保由勝任的人員審閱及批准每項輸出,方予依賴、向僱員公佈、用於薪酬或申報,或用於任何重大影響個人的決定。客戶不得在未經有意義的人手審閱下,使用 AI 功能作出僱傭決定(包括招聘、影響薪酬或休息的排更、紀律處分、降職或解僱)。

3. 輸出的擁有權

3.1 輸入。 客戶(或其授權人)保留輸入的一切權利。客戶授予 SlashHub 一項許可,僅為提供 AI 功能及按 DPA 允許而處理輸入。

3.2 輸出 — 擁有權。 在 SlashHub 與客戶之間,客戶擁有輸出(在可擁有的範圍內)。SlashHub 將其在為客戶產生的輸出中可能擁有的所有權利轉讓予客戶,並不主張對該輸出的所有權。

3.3 不保證可保護性或唯一性。 鑑於生成式 AI 的性質,SlashHub 不保證輸出可受版權或其他知識產權保護、屬唯一,或不與為他人產生的輸出或現有第三方材料相似。客戶須負責評估輸出是否適合使用、保護或註冊。

3.4 向客戶授予的許可。 在付款及遵守的前提下,SlashHub 授予客戶一項全球、非獨家、免版稅、可撤銷的許可,為客戶的內部業務目的使用、複製、修改及分發輸出。本許可不准許使用輸出訓練競爭 AI 模型,或在會構成誤導或法律要求披露的情況下,將 AI 輸出冒充為人類所創作。

4. 模型訓練權限

4.1 我們不使用客戶資料訓練自身模型。 SlashHub 不使用客戶資料(包括僱員個人資料)訓練、微調或改進其自身的任何 AI 模型,亦不售賣客戶資料。此與 DPA 一致。

4.2 模型供應商。 SlashHub 與模型供應商訂約,條款(a)禁止彼等使用客戶資料訓練其通用模型,及(b)施加不低於 DPA 的保密及保安義務。如某模型供應商就某功能無法符合該等條款,SlashHub 將不會將個人資料路由至該功能。

4.3 匿名化改進(選擇加入)。 SlashHub 可使用無法合理連結至任何個人的匯總、匿名化資料,以營運、保護及改進服務。如客戶希望退出任何可選的匿名化改進計劃,可於帳戶設定中作出,或電郵 ai@slashhub.hk。

4.4 客戶本身的訓練資料。 客戶不得使用 AI 功能訓練、微調或蒸餾競爭模型,亦不得提交其無權提交的輸入。

5. 人手監督及透明度

5.1 每項不可逆的 AI 驅動操作(例如建立記錄、發送訊息或匯出薪酬)均需客戶或授權使用者確認,或會被記錄並可逆。

5.2 AI 的參與會在介面中標示,AI 工具呼叫會記錄於帳戶活動日誌。如客戶發佈或分享輸出,客戶須負責在法律要求時披露其 AI 來源。

6. 準確性、核實及責任

6.1 不保證準確性。 SlashHub 不就任何輸出的準確性、完整性、時效性或對特定用途的適用性作出保證。客戶必須在依賴任何輸出前獨立核實,尤其是薪酬、強積金、法定假日、假期、超時、稅務及 IR56 數字。

6.2 責任。 SlashHub 就 AI 功能的責任,受 MSA 第 12 條的除外及上限(及 MSA 第 12.3 條的例外)規限。SlashHub 對任何人依賴輸出而作出的決定,或對因未按第 6.1 條核實的輸出所產生的任何損失,概不負責。

6.3 客戶彌償。 客戶應就因客戶使用輸出(包括在未經第 2.2 及 5.1 條所要求的審閱下作出的自動化決定或僱傭決定)而產生的申索,彌償 SlashHub。

7. 歐盟 AI 法及受規管用途

7.1 SlashHub 將 TimePlate AI 功能分類為歐盟 AI 法下的有限風險,並依賴透明度及人手監督(本附件及 AI 披露聲明)。該等功能並非擬作被禁止的 AI 用途,SlashHub 亦不擬將其用於生物特徵分類或情緒識別。

7.2 僱傭場景。 客戶確認,在僱傭場景中使用的 AI,可於歐盟 AI 法(附件三)下屬高風險。如客戶於歐洲經濟區設立或針對歐洲經濟區的個人,客戶須負責任何適用的額外義務,包括人手監督、資料治理、註冊及任何基本權利影響評估,並須相應配置 AI 功能(輔助性、經人手審閱)。SlashHub 僅以輔助工具形式提供 AI 功能,並不保證符合客戶的行業特定義務。具有受規管或高風險用途的客戶應聯絡 ai@slashhub.hk。

8. 測試版 AI 功能

標記為 beta/preview/experimental 的 AI 功能按「原樣」提供,可能變更或撤回,不附服務水平承諾,且客戶不應在未經自行測試下將其用於真實薪酬或其他具重大後果的資料。

9. 期限、變更及聯絡

9.1 本附件隨 MSA 生效,並受其管轄法律及爭議解決條款規限(香港法律;香港國際仲裁中心仲裁,仲裁地為香港)。

9.2 SlashHub 可按 MSA 第 19 條更新本附件,重大變更須提前 30 日通知。

9.3 AI 查詢及舉報:ai@slashhub.hk。

SlashHub Limited © 2026 · For questions, contact legal@slashhub.hk