Contract for the use of the AI services of audiotranskription

hereinafter referred to as the "Usage Agreement".

The contract is concluded between the user and:

audiotranskription -- dr. dresing & pehl GmbH

Deutschhausstraße 22a

35037 Marburg

Germany

Managing Directors: Dr Thorsten Dresing and Thorsten Pehl

Marburg Local Court HRB 5171

Telephone: +49 6421 590979-0

Email: info@audiotranskription.de

Hereinafter referred to as "audiotranskription"

1. Subject matter of the contract

1.1. The subject of this contract is the provision of the following AI-supported services within the f4 software by audiotranskription:

a) Automatic transcription (hereinafter: "transcription"): The automatic conversion of audio data into a text file by means of an AI-supported speech recognition model. The supported languages are German and English; any further languages can be found in the product description (audiotranskription.de/spracherkennung).

b) AI-supported analysis services (hereinafter: "AI Services"): The transmission of text data to a language model (Large Language Model,
LLM) in order to obtain AI-generated interpretative suggestions, descriptions and methodologically reflective insights. The AI Services are provided using the infrastructure of the Gesellschaft für wissenschaftliche Datenverarbeitung mbH Göttingen (GWDG) or of STACKIT GmbH & Co. KG (STACKIT). Both providers are used as equivalent sub-processors; in both cases, processing takes place exclusively on servers located in the EU.

1.2. Both services utilise AI technology but differ in the nature and scope of the processing: transcription converts audio data into text by means of an AI-supported speech recognition model, whereas the AI Services transmit text data selected by the user to a language model
(LLM) for the generation of interpretative suggestions.

1.3. Transcription and the AI Services may be used independently of one another or in combination. The AI Services are an optional additional service. The user decides independently and actively which service is to be used. There is no automatic transfer of data between the services --- in particular, audio data is not automatically transmitted to the language model (LLM), nor is text data automatically transmitted to the speech recognition system.

1.4. The functionality of the transcription is described in Clause 3. The functionality of the AI Services is described in Clause 4.

1.5. In order to use the services, in addition to acceptance of these contractual terms, the conclusion of a data processing agreement is generally required as part of the user's registration. It is the user's own responsibility to assess whether the conclusion of a data processing agreement is necessary. The basic requirements for the processing of personal data are described in the data processing agreement and in the privacy policy on the website (https://www.audiotranskription.de/Datenschutz). The user may make individual additions, e.g. regarding the types of personal data.

1.6. It is the user's responsibility to point out the need for an additional non-disclosure agreement. Where this is necessary for the user's activities, audiotranskription will provide an additional confidentiality agreement in accordance with Section 203 of the German Criminal Code (StGB).

1.7. In order to use the services within the f4 software, the user must have created a personal account.

1.8. For transcription, quotas of time units may be credited to the account set up by the user. The time units must be purchased separately in advance by the user via the web shop (https://www.audiotranskription.de/shop/). The purchase of the time units required for transcription is subject solely to the contractual terms of the web shop. Transcription is commissioned by way of individual projects in accordance with Clause 5 and is billed in time units.

1.9. The use of the AI Services is included in the licence for the f4 software. A separate purchase of quotas is not required. The AI Services are used by way of individual enquiries in accordance with Clause 6. The use of the AI Services is subject to a fair use policy: in order to prevent server overload and to ensure equal access for all users, the software limits the number of enquiries within a given period. Once the enquiry limit has been reached, usage is temporarily paused; once the pause has ended, further enquiries may be made.

1.10. The AI Services are activated by the user within the f4 software by logging in with an f4 account. Without this activation, no processing of data by the AI Services takes place.

1.11. Access to the Internet is not covered by this contract. The user bears sole responsibility for the functionality of the Internet access, including the transmission paths of the computer used.

1.12. The conditions for transcription are set out in Clause 9. The use of the AI Services is included in the software licence and is subject to the fair use policy in accordance with Clause 1.9.

1.13. The rights and obligations of the parties arise solely from the provisions of this contract and any additional data processing agreement concluded, as well as the obligation to maintain confidentiality in accordance with Section 203 StGB. General terms and conditions of the user shall not become part of the contract, even if audiotranskription does not expressly object to them. In the event of any conflict between the provisions of this contract and the provisions of the data processing agreement and/or the obligation of confidentiality pursuant to Section 203 StGB, the provisions of the latter two documents shall take precedence within their scope of application.

2. Amendments to the contractual provisions

2.1. audiotranskription reserves the right to amend these contractual provisions, in particular as a result of further developments in the services, even within an existing contractual relationship.

2.2. The user shall be informed of any planned changes to the contractual provisions by email at least 6 weeks before the changes come into effect. In this email, the changes to the previous contractual provisions will be identified and, where necessary, explained in addition.

2.3. The user may object to the changes to the contractual provisions within 6 weeks of receipt of the aforementioned email. If the user does not object to the changes, or does not do so in good time, their consent to the amended contractual provisions shall be deemed to have been granted and the changes shall be deemed to have been effectively agreed. The user will be informed of this legal consequence separately in connection with the planned changes.

3. Functionality of the transcription

3.1. The transcription converts audio files into writing by means of an AI-supported speech recognition model. The speech recognition is trained for standard High German. Further training, languages and specialisations can be found in the product description (audiotranskription.de/spracherkennung).

3.2. According to the tests carried out by audiotranskription and the resulting evaluation, automated transcription can be recommended for individual interviews with good recording quality. In some settings, however, manual transcription may be faster. Details of the aforementioned speech recognition test results can be found on the audiotranskription website at https://www.audiotranskription.de/spracherkennung.

3.3. An installed trial version, demo version, a browser client or a licensed version of f4, from version 2024 onwards, is required in order to use the transcription. A licence for the f4 software is not part of this contract and must be purchased separately. An Internet connection is required to use the transcription in order to upload media files and download the text files.

4. Functionality of the AI Services

4.1. The AI Services enable the user, within the f4 software, to transmit selected text excerpts (e.g. from transcripts, field notes or coded passages) together with a question or instruction (a "prompt") to a language model (LLM) and to receive an AI-generated response in return.

4.2. The AI Services are designed as a methodological tool to support the qualitative and interpretative analysis of text data. They do not provide ready-made analyses, but rather interpretative suggestions, inspiring ideas and reflective questions. The AI serves as a "sparring partner" that helps users to question their own interpretations, discover new perspectives and increase the depth of their analysis.

4.3. The AI Services specifically support scientific methods such as:

• "Hybrid Interpretation": Here, the dialogue with the AI simulates an interpretative discussion group. The texts generated by the AI serve as alternative readings and hypotheses, which the user critically examines, rejects or integrates into their own analysis.
• "Serendipity Prompting": Here, the AI's ability to establish unexpected and creative cross-connections is utilised. The aim is to guide the user, through surprising questions or associations, towards new, insight-promoting leads within the data material that would have remained hidden through a linear approach.

4.4. The authority over the interpretation of content, the methodological control and the final analytical output remain entirely and unrestrictedly with the user at all times. The AI is merely a tool for inspiration and reflection.

4.5. The AI Services use a language model operated, at audiotranskription's choice, on servers belonging to the GWDG (Gesellschaft für wissenschaftliche Datenverarbeitung mbH Göttingen) or to STACKIT GmbH & Co. KG (STACKIT). The GWDG is a joint institution of the Georg-August University of Göttingen and the Max Planck Society. STACKIT is a cloud service provider of the Schwarz Group based in Neckarsulm. Both providers are used as equivalent sub-processors; which provider processes a specific enquiry depends on the respective configuration and availability. In both cases, processing takes place exclusively on servers located in the EU.

4.6. Processing within the scope of the AI Services is designed to be transient: the transmitted text data is deleted immediately after the response has been generated. No order data remains with audiotranskription or its sub-processors.

4.7. audiotranskription reserves the right to change or update the language model used at any time (e.g. model changes, version updates). The models used in each instance, as well as the prompts employed, can be viewed transparently by the user via the software's export function.

4.8. A change of the provider of the AI infrastructure (sub-processor) shall only take place with at least 14 days' prior notice by email or via the software. The user has the right to object to the change of provider within this period. In the event of a timely objection, the user may deactivate the AI Services with immediate effect.

5. Commissioning of an individual project for transcription

5.1. For a transcription, the user uploads audio data using the software via a TLS-encrypted connection through the relay server of audiotranskription to the speech recognition server. The transcription supports the following file formats for the audio data: mp3, m4a, wav, aiff. Further formats may be added according to the state of the art.

5.2. The output of the transcription result, the text file, is provided in rtf format. Further formats may be added according to the state of the art.

5.3. When an audio file is uploaded to the server, audiotranskription automatically checks, on the basis of the size of the file, whether sufficient time quotas are stored in the account. If no or insufficient time quotas are stored, transcription is not possible and the software displays a corresponding error message. Provided that sufficient time quotas are stored, the transcription is carried out and the service rendered is billed via the stored quotas in accordance with Clause 9. The user can download the transcribed text file via the user interface and save it locally on their device.

5.4. audiotranskription shall commence the transcription, where the user is a consumer within the meaning of Section 13 of the German Civil Code (BGB), as soon as the user has expressly consented to the commencement of the provision of services before the end of the withdrawal period. By giving their consent, the user must at the same time expressly confirm that they are aware that, upon commencement of the provision of services by audiotranskription, they lose their right of withdrawal with regard to the specific individual project, the transcription contract and all further or subsequent individual projects based on this contract.

6. Use of the AI Services (individual enquiries)

6.1. For an individual enquiry, the user selects a text excerpt in the f4 software and actively submits it. The data is forwarded via a TLS-encrypted connection through the relay server of audiotranskription to the sub-processor used in the respective case (GWDG or STACKIT).

6.2. The sub-processor processes the transmitted data exclusively for the purpose of generating a response to the specific enquiry.

6.3. The generated response is sent back to the user's software and displayed there. The entire dialogue is stored exclusively locally by the user.

6.4. With each enquiry, the software automatically checks whether the user is authorised to use the AI Services and whether the enquiry limit under the fair use policy (see Clause 1.9) has not yet been reached. Without valid authorisation, or once the enquiry limit has been reached, a corresponding message is displayed and the enquiry is not processed.

6.5. The user can export the AI dialogue at any time and save it locally. The user is solely responsible for the retention and protection of the locally stored dialogue data.

7. Obligations of the user

7.1. The user must independently assess whether the conclusion of a data processing agreement is required (see Clause 1.5) and/or an additional confidentiality agreement pursuant to Section 203 StGB is to be concluded (see Clause 1.6).

7.2. The user undertakes not to upload, or transmit to the AI Services, any unlawful content that violates the law, official regulations or the rights of third parties.

7.3. The user must protect the access data to the account and all confidential information from access by unauthorised third parties. Should third parties nevertheless have gained unauthorised access, or at least should the possibility of this exist, the user shall inform audiotranskription without delay.

7.4. Transcription: The selection of suitable audio data for transcription is the sole responsibility of the user. Dialects, background noises, changing speakers, poor recording quality and other background noises regularly have a negative effect on recognition accuracy and therefore on the transcription result. The more of the aforementioned aspects appear in the audio file, the worse the transcription result. As a rule, unsuitable audio files lead to words, sentences and/or passages that have not been transcribed or have been transcribed incorrectly. In individual cases, the entire transcription result may even be unusable. Where the recording is not available in an explicitly supported language, error-free transcription is altogether impossible. Where the transcription result is based on an inadequate selection of suitable audio data, this shall have no effect on audiotranskription's claim to full remuneration, regardless of the result of the transcription.

7.5. AI Services: The selection of the text data to be transmitted to the AI Services is the sole responsibility of the user. The processing of special categories of personal data within the meaning of Article 9 of the GDPR (e.g. health data, political opinions, religious beliefs) is excluded from the AI Services. The user is obliged to ensure, prior to each transmission, that no such data is contained therein, and to consider and implement appropriate measures to protect the data subjects.

7.6. AI Services: The user acknowledges that the AI Services are based on language models which may, in principle, generate incorrect, incomplete, outdated or misleading responses (so-called "hallucinations"). The responses of the AI do not constitute scientifically validated results, statements of fact or legal advice. It is the user's responsibility to critically examine and verify all AI-generated content and to decide, on their own responsibility, on its use.

7.7. The user shall use a state-of-the-art virus protection programme and, in particular, check the audio data for viruses before sending it, as well as ensure that the f4 software used is kept up to date.

8. Scope of services

Transcription:

8.1. audiotranskription endeavours to provide the transcription service 24 hours a day. Nevertheless, outages are possible, in particular due to maintenance work.

8.2. Provided that the user makes the audio file available in a suitable file format in accordance with Clause 5.1, audiotranskription will transcribe the audio file into a text file fully automatically using the AI-supported speech recognition model. The data transfer takes place --- as with the AI Services --- via a TLS-encrypted connection through the relay server of audiotranskription.

8.3. Within the scope of the fully automated transcription, audiotranskription has no knowledge of the content of the audio file provided. In this respect, audiotranskription will in particular: not check the suitability (see Clause 7.4) of the audio file provided by the user, not carry out any quality assurance in connection with the transcription, and not take into account any individual specifications of the user.

8.4. Since the speech recognition, as described in Clause 3, does not function without error, and the transcription result also depends on the suitability of the audio file (see Clause 7.4), the quality of the transcription result may vary depending on the individual case. Accordingly, it must be assumed that the transcription result is, as a rule, not free of errors. In this respect, audiotranskription does not guarantee that the transcription result is free of errors, but only that the speech recognition has the functionalities described in Clause 3. Any correction of errors or correction of the transcription result is therefore the sole responsibility of the user.

8.5. Individual post-editing or error correction of transcription results by audiotranskription is not owed.

AI Services:

8.6. audiotranskription endeavours to provide the AI Services 24 hours a
day. Nevertheless, outages are possible, in particular due to maintenance work, updates to the language model or unavailability of the sub-processor's infrastructure.

8.7. audiotranskription transmits the enquiries submitted by the user to the language model fully automatically and returns the generated response to the user's software.

8.8. Within the scope of the fully automated processing, audiotranskription has no knowledge of the content of the transmitted text data or of the generated responses. In this respect, audiotranskription will in particular: not check the substantive suitability of the transmitted text data, not carry out any quality assurance in connection with the AI-generated responses, and not take into account any individual specifications of the user.

8.9. Since language models do not, in principle, operate without error, and the quality of the responses depends on numerous factors (including the wording of the prompt, the quality and scope of the transmitted text data, and the current model version), the quality of the AI-generated responses may vary considerably depending on the individual case. audiotranskription does not guarantee the substantive correctness, completeness or scientific validity of the AI-generated responses. The responses expressly do not constitute finished research results, but serve to stimulate and reflect.

8.10. Individual adaptation, post-editing or correction of AI-generated responses by audiotranskription is not owed.

9. Remuneration and payment

Transcription: For each transcription, audiotranskription shall receive remuneration in the form of time quotas subject to the following conditions: time quotas can be purchased by the hour. For one hour of interview, EUR 19.99 incl. VAT is charged. Volume discounts and discounts for special groups of people (e.g. students) are possible. Billing is per minute or part thereof. Test versions that are marked as such may, within the scope of their technical possibilities and restrictions, also be used free of charge where applicable.

AI Services: The use of the AI Services is included in the licence for the f4 software and is not remunerated separately. Use is subject to the fair use policy in accordance with Clause 1.9.

10. Rights to the results

10.1. Transcription: audiotranskription grants the user an exclusive right of use, unlimited in time, space and content, to the transcription results obtained under this contract and in connection with the respective individual projects.

10.2. AI Services: The user may use, edit and publish the responses and texts generated within the scope of the AI Services without restriction for their own purposes.

10.3. audiotranskription does not assert any rights to the data transmitted by the user (audio data, text data, prompts) or to the results generated (transcripts, AI-generated responses). The transmitted data remains the sole property of the user.

10.4. The user acknowledges that, under the current legal situation, AI-generated texts generally do not enjoy copyright protection, as they are not regarded as a personal intellectual creation within the meaning of Section 2 of the German Copyright Act (UrhG). The question of eligibility for protection must be assessed on a case-by-case basis.

10.5. Where the user uses AI-generated content in scientific publications or other publications, it is their responsibility to comply with the applicable citation and disclosure rules of the respective professional community, journal or institution, and to transparently indicate the AI support.

11. Data protection / costs for cooperation services in connection with the data processing agreement

11.1. The parties undertake to comply with the relevant statutory provisions on the handling of personal data, in particular the German Federal Data Protection Act (BDSG) and the EU General Data Protection Regulation (GDPR).

11.2. In order to enable the processing of personal data within the scope of the services, the conclusion of a data processing agreement is generally required. This can be concluded as part of the registration process or subsequently in the customer area and can be viewed at audiotranskription.de/AVV.

11.3. The user is advised that the use of the services --- in particular the AI-supported analysis services --- for qualitative research data may require a data protection impact assessment (DPIA) in accordance with Article 35 of the GDPR. The obligation to assess and, where applicable, carry out a DPIA rests with the user. Upon request, audiotranskription shall provide the information required for this purpose regarding the nature of the processing, the technical and organisational measures and the sub-processors used.

11.4. Where individual instructions go beyond the provisions of this contract and/or the data processing agreement and require additional work on the part of audiotranskription, these shall require the prior consent of audiotranskription and shall be remunerated separately in accordance with the price list valid at the relevant time.

12. Liability

12.1. Unless otherwise agreed in this contract, the parties shall be liable in accordance with the statutory provisions.

12.2. Transcription: audiotranskription does not guarantee that the transcription result is free of errors (see Clause 8.4).

12.3. AI Services: audiotranskription is not liable for the substantive correctness, completeness or suitability of the AI-generated responses. The use of the AI-generated content is at the user's own responsibility.

12.4. AI Services: audiotranskription is not liable for damages arising from the user using AI-generated responses, without their own examination, as statements of fact, scientific results or a basis for decisions.

12.5. AI Services: audiotranskription is not liable for the temporary unavailability of the AI Services, insofar as this is due to circumstances outside the control of audiotranskription, in particular disruptions to the sub-processor's infrastructure.

12.6. In the event of any claim by a third party, the user is obliged to provide audiotranskription, free of charge and without delay, with all information available to them concerning the matter in question.

13. Confidentiality

13.1. The parties undertake to maintain confidentiality regarding confidential information. "Confidential information" is in particular trade secrets within the meaning of Section 2 No. 1 GeschGehG, as well as all information and documents of the other party that are marked as confidential or are to be regarded as confidential on the basis of the circumstances, in particular interview data, transmitted text data, prompts, AI-generated responses, operational processes and other know-how.

13.2. Exempt from this obligation is such confidential information as was demonstrably known to the parties at the time the contract was concluded or subsequently becomes known to them from a third party, without thereby violating a confidentiality agreement, statutory provisions or official orders; was publicly known at the time the contract was concluded or is subsequently made public, unless this is due to a breach of this contract; or which must be disclosed on the basis of statutory obligations or by order of a court or authority. As far as permissible or possible, the parties shall inform each other in advance and give each other the opportunity to take action against the disclosure.

13.3. The parties shall only grant access to confidential information to advisers who are subject to professional secrecy or who have previously been subject to obligations corresponding to the confidentiality obligations of this contract.

13.4. Taking into account the state of the art and --- insofar as personal data is concerned --- specifically in accordance with Article 32 of the GDPR, the parties shall take appropriate technical and organisational measures to protect confidential information and personal data, and shall implement a procedure for the regular review, assessment and evaluation of the effectiveness of the technical and organisational measures to ensure the security of the processing.

14. Duration of contract / termination

14.1. The contractual relationship ends with the termination of the customer account. The customer account can be terminated with a notice period of two weeks, whereby the following conditions apply in the event of termination:

14.2. If the user terminates the contract even though time quotas (transcription) are still stored in the relevant account, the quotas shall expire at the end of the contract term. audiotranskription shall not pay any compensation for this.

14.3. audiotranskription may, except in the case of an extraordinary reason for termination, terminate the contract with two weeks' notice if, at the time of termination, no quotas are stored in the relevant account. Where quotas are stored in the account, audiotranskription may, subject to the notice period set out in Clause 14.1, however terminate the contractual relationship at the (next possible) point in time at which the quota stored at the time of the notice of termination has been used up.

14.4. The AI Services may be terminated by the user at any time by deactivation within the f4 software, without this affecting the contractual relationship in other respects. The deactivation takes effect immediately; from that point onwards, no further AI enquiries are processed.

14.5. audiotranskription may discontinue or substantially change the AI Services with four weeks' notice to the end of the month, in particular where the providers used as sub-processors (GWDG or STACKIT) cease to provide the infrastructure or where the legal framework no longer permits continuation.

14.6. The right to extraordinary termination for good cause remains unaffected for both parties.

14.7. To be effective, a notice of termination must be given in text form. The user's notice of termination must be addressed to info@audiotranskription.de.

15. Right of withdrawal

15.1. Where the user is a "consumer" within the meaning of Section 13 BGB, they have, taking into account the provisions of Clause 5.4, the right to withdraw from this contract within fourteen days without giving any reason.

15.2. The withdrawal period is fourteen days from the date of conclusion of the contract.

15.3. In order to exercise the right of withdrawal, the user must inform audiotranskription (audiotranskription -- dr. dresing & pehl GmbH, Deutschhausstr. 22a, 35037 Marburg, Germany, Fax: +49 6421 590979-20, Email: info@audiotranskription.de) of their decision to withdraw from this contract by means of a clear statement (e.g. a letter sent by post, fax or email). For this purpose, they may use the attached model withdrawal form, although this is not mandatory.

15.4. In order to comply with the withdrawal period, it is sufficient for the user to send the notification of the exercise of the right of withdrawal before the expiry of the withdrawal period.

16. Consequences of withdrawal

16.1. If the user withdraws from this contract, audiotranskription shall reimburse the user for all payments that audiotranskription has received from them, including delivery costs (with the exception of the additional costs resulting from the user having chosen a type of delivery other than the cheapest standard delivery offered by audiotranskription), without delay and at the latest within fourteen days of the day on which the notification of the withdrawal from this contract is received by audiotranskription. For this reimbursement, audiotranskription will use the same means of payment that the user used for the original transaction, unless expressly agreed otherwise; in no case will fees be charged for this reimbursement.

16.2. Where the user has consented, in accordance with Clause 5.4, that the service should commence during the withdrawal period, the user shall lose their right of withdrawal with regard to the specific individual project, as well as all further or subsequent individual projects based on this contract. Where the user has consented that the AI Services should be provided before the expiry of the withdrawal period, they shall pay audiotranskription an appropriate amount corresponding to the proportion of the services already provided up to the time of the withdrawal compared to the total scope of the agreed service.

17. Model withdrawal form

If the user wishes to withdraw from the contract, they should complete this form and return it. The use of this form is not mandatory.

To audiotranskription -- dr. dresing & pehl GmbH, Deutschhausstr. 22a, 35037 Marburg, Germany

Fax: +49 6421 590979-20, Email: info@audiotranskription.de:

I/we (*) hereby withdraw from the contract concluded by me/us (*) for the provision of the following services

Ordered on (*) / received on (*)

Name of the consumer

Address of the consumer

Signature of the consumer (only for notification on paper)

Date

(*) Delete as appropriate.

18. Destruction or deletion of confidential information after termination of the contract

audiotranskription is obliged, in the event of termination of the contract, to destroy or delete all information, in particular the access data to the account, in accordance with data protection regulations. Exempt from this is data for which a statutory retention period exists. With regard to the AI Services, the following applies in addition: as the AI Services are designed as ephemeral processing, no order data remains with audiotranskription or its sub-processors after the completion of the respective individual enquiry.

19. Online dispute resolution

The European Commission provides a platform for out-of-court online dispute resolution (ODR platform), which can be accessed via the external link www.ec.europa.eu/consumers/odr. We endeavour to settle any disputes arising from or in connection with this contract by mutual agreement. Beyond this, we are not obliged to participate in arbitration proceedings and do not offer to participate in such proceedings.

20. Applicable law and place of jurisdiction (only for entrepreneurs)

20.1. German law shall apply to this contract.

20.2. The exclusive place of jurisdiction for all disputes arising from or in connection with this contract shall be the registered office of audiotranskription. audiotranskription is also entitled to sue the user at the user's registered office.