AI Disclaimer, CallMerlin

This AI Disclaimer ("Disclaimer") forms an integral part of the General Terms and Conditions of CallMerlin and applies to all AI services offered through the Platform.

The Platform offers the following AI services: a) AI voice assistant: an AI-powered voice assistant that handles inbound telephone calls and conversations via the embeddable voice widget. b) Speech recognition and transcription: automatic conversion of speech to text. c) Intent recognition: automatic interpretation of what a Caller or PIN User requests. d) Action execution: automatic execution of actions in Connected Services (such as creating appointments, tickets or records) based on the conversation. e) AI summaries: automatic summaries of calls and their outcomes.

2.1. AI services are based on statistical language and speech models. Output is generated based on patterns in training data and is not the result of human judgment or fact-checking. 2.2. CallMerlin does not guarantee that AI-generated output: a) is factually correct; b) is complete; c) is free from biases; d) is suitable for a specific purpose; e) correctly reflects what a Caller said or intended. 2.3. The User is at all times responsible for reviewing, verifying, and approving AI-generated output and the actions performed by the voice assistant, before relying on them or making them available to third parties.

3.1. AI services provide informational output and do not constitute a substitute for professional advice. AI output shall not be construed as: a) legal advice; b) medical or health advice; c) financial or investment advice; d) tax advice; e) accounting advice; f) psychological or therapeutic advice. 3.2. The User who uses AI output in a context where professional advice is required does so entirely at their own risk and is advised to consult a qualified professional.

4.1. Transparency: The User is obligated to inform Callers that they are speaking with an AI system. This is a requirement under the EU AI Act (Regulation (EU) 2024/1689, Article 50). CallMerlin provides a standard announcement message for this purpose. 4.2. Transcription: Calls are automatically transcribed for functional purposes (such as scheduling appointments and executing actions) and quality improvement. Audio files are not permanently retained. Transcripts are retained in accordance with the retention periods in the Privacy Policy. 4.3. Limitations: The AI voice assistant may make errors in: a) understanding accents, dialects, or background noise; b) scheduling appointments (wrong date, time, or details); c) executing actions in Connected Services (wrong data or wrong action); d) understanding complex or ambiguous requests. 4.4. Liability: CallMerlin is not liable for damage arising from incorrect or missed actions by the AI voice assistant, including missed appointments, incorrectly relayed information, incorrectly executed actions in Connected Services, or failure to transfer urgent calls. 4.5. No emergency calls: the voice assistant and the telephone numbers provided are not suitable for emergency calls (such as 112 or 911). The User shall ensure an alternative means of contacting emergency services. 4.6. Do-not-call register: For outbound calls, the User is responsible for compliance with applicable do-not-call registers and all regulations regarding unsolicited telephone communications. 4.7. Availability: The AI voice assistant may be unavailable due to maintenance, outages, or capacity constraints. CallMerlin offers no guarantee of 100% availability.

5.1. The voice assistant can perform actions in Connected Services based on what a Caller or PIN User requests. The interpretation of spoken instructions is probabilistic by nature; misinterpretation cannot be ruled out. 5.2. The User determines, through the configuration of connections and permissions, which actions the voice assistant may perform. The User is advised to grant only the minimally required permissions and to review executed actions regularly via the call logs. 5.3. CallMerlin does not guarantee that actions in Connected Services are executed correctly, completely or on time, and is not responsible for the processing of such actions by the Connected Service itself.

6.1. CallMerlin uses AI models from xAI, Inc. (Grok) as sub-processor for delivering AI services (chat, voice, speech recognition and language understanding). 6.2. CallMerlin exclusively contracts with AI providers that: a) provide a data processing agreement; b) offer an opt-out for using customer data for model training; c) implement appropriate security measures. 6.3. Personal data processed in AI requests is not used for training AI models. 6.4. CallMerlin may switch AI providers without prior notice, provided the new provider offers at least the same security and privacy safeguards.

7.1. The intellectual property status of AI-generated works under Dutch and European law has not been fully settled. There is no certainty that AI-generated content enjoys copyright protection. 7.2. CallMerlin makes no claim to AI-generated output produced at the User's request. 7.3. The User accepts that: a) similar AI output may also be generated for other users; b) exclusivity of AI-generated output cannot be guaranteed; c) AI output may inadvertently resemble existing works of third parties. 7.4. The User indemnifies CallMerlin against third-party claims regarding intellectual property rights in AI-generated output used or published by the User.

8.1. CallMerlin endeavors to comply with the European Artificial Intelligence Act (EU AI Act, Regulation (EU) 2024/1689), insofar as applicable. 8.2. CallMerlin's AI services are classified as limited risk within the meaning of the AI Act. The transparency obligations of Article 50 AI Act apply: data subjects must be informed that they are interacting with an AI system. 8.3. The User, as deployer within the meaning of the AI Act, is responsible for: a) informing Callers and other data subjects that they are communicating with an AI system; b) monitoring AI output and intervening when undesirable results occur; c) documenting the use of AI systems insofar as legally required.