WizBI AI Use Policy

1. What the Service Is

1.1. Decision-support tool. The Service uses language and speech models to interpret questions, retrieve and analyse data, summarise, suggest and alert. It is an aid to decision-making. It is not professional advice of any kind, including business, accounting, tax, legal, investment, financial, insurance, medical, veterinary, agronomic, engineering, safety or employment advice, is not tailored to any individual’s circumstances, and is no substitute for a qualified professional. The Service is not a party to any transaction, decision, engagement or report the Customer makes in reliance on it. The Service is a business-specific tool that operates on the Customer’s own data and configuration; it is not a general-purpose AI assistant. (Section 10.1 of the Terms.)

2. Human Review

2.1. The Customer verifies. The Customer is responsible for reviewing every Output, and for professional human review and verification against the source before relying on an Output for a decision of consequence, and in any event before: a payment, commitment or transaction; an accounting entry or a report to an authority, tax office, bank, auditor or board; a decision about an employee, candidate, customer or supplier; a health or treatment decision (including for animals); and a safety action. When the Customer passes an Output to others it will tell them it was produced by AI and should not be relied on without review. (Section 10.3 of the Terms.)

3. Transparency

3.1. AI disclosure. Where the channel allows, Users are informed that they are interacting with an AI system: the Service is designed to display, in its channels, at the start of a conversation or on joining, a notice that the User is conversing with an AI system, and to announce in a voice call that the call is with an AI and whether it is recorded (Section 10.4 of the Terms). This is the transparency the Service provides for the purposes of Article 50 of the EU Artificial Intelligence Act (Regulation (EU) 2024/1689), where that Act applies.

3.2. Generated media. Where the Service generates synthetic audio, image or video content, WizBI marks it, or uses a model that marks it, as AI-generated in a machine-readable form where technically feasible (Section 10.4 of the Terms).

3.3. The Customer’s part. Where a channel does not allow the notice, the Customer gives it to its Users. The Customer will not remove, hide or alter the notice, will not present an Agent as a person, and is responsible for any further disclosure the law requires of it, including when it publishes AI-generated content (Section 10.4 of the Terms).

4. High-Risk Uses

4.1. High-risk uses. The Customer will not use the Service to make, automatically and without review and approval by a qualified person, a decision with legal or similarly significant effect on an individual, including in employment (recruiting, screening, evaluation, promotion, dismissal, pay), credit and credit scoring, insurance and underwriting, housing, education and student assessment, eligibility for public services or benefits, health, diagnosis and treatment, justice, law enforcement and immigration, and critical infrastructure. A Customer that uses the Service in one of these fields, even with a human in the loop, will tell WizBI in advance, is responsible for every regulatory obligation that applies to it (including as a deployer under the EU AI Act, where it applies), and indemnifies WizBI under Section 15.1 of the Terms. (Section 10.7 of the Terms.)

5. Prohibited AI Uses

5.1. Prohibited AI uses. In addition to the Acceptable Use Policy, the Service must not be used for: emotion recognition of employees or students; biometric categorisation of people by sensitive characteristics; facial or voice recognition for surveillance; social scoring; manipulation exploiting a person’s vulnerabilities; sexual content involving minors or intimate content without consent; impersonating a real person or creating a deepfake without consent and labelling; developing weapons or malware or attacking infrastructure; fraud, deception, spam or disinformation; licensed professional advice to the public without review by a licence holder; and any use prohibited by the relevant model provider’s usage policy. (Section 10.8 of the Terms.)

6. Voice and Biometrics

6.1. Voice and biometrics. Recordings and media submitted to the Service are processed for transcription, understanding and response and kept as set out in Section 9.8 of the Terms and the DPA. The Service does not create voiceprints and does not identify people by voice or face unless expressly ordered and with the required consent. Cloning a person’s voice requires that person’s written consent and is the Customer’s responsibility. (Section 10.9 of the Terms.)

7. Enforcement, Changes and Language

7.1. Enforcement. A breach of this Policy is a breach of Section 10 of the Terms, and WizBI’s rights and remedies under the Terms apply to it, including suspension under Section 16.4 and termination under Section 16.5 of the Terms. Agent actions are governed by Section 10.5 of the Terms.

7.2. Changes. This Policy is updated, and its versions are archived at the Legal Hub, under Section 17 of the Terms.

7.3. Language. This Policy is made in English, which governs. A Hebrew or other translation is for convenience only.