DNV AI Assistant
Terms of use
Terms of use
Your DNV AI Assistant is built to guide customers through their digital journey with DNV. Powered by generative artificial intelligence (GenAI) and trained on publicly available DNV content, the AI Assistant is a beta version and is not capable of providing authoritative or professional advice.
Trained on public DNV website content, including publications, news, tools and services content, the AI Assistant – currently in beta version – is meant to help visitors on the DNV website find information about DNV services and resources and receive suggestions for content to visit next.
The AI Assistant comes with limitations. While ultimately intended to support all sectors, the assistant is currently configured to interact using content related to the Maritime industry and therefore is made available on Maritime related pages.
As a generative AI powered assistant, the application is configured and intended to support, assist and provide direction, but technical inherent limitations of generative AI make the AI Assistant susceptible to errors (so-called hallucinations). Owing to this, users are strongly advised not to take any critical decisions based on its responses alone and to always validate the information received.
DNV’s Vision is to be a trusted voice to tackle global transformations. Our business is built on trust, and our values reflect who we are as an organization. Our policies, principles and guidelines shall ensure that DNV’s services and activities are carried out in compliance with applicable laws and regulations, including but not limited to the EU AI Act and GDPR.
In the context of AI applications DNV is amongst others committed to:
These terms and conditions shall be governed by and construed exclusively in accordance with the laws of Norway, without regard to principles of conflicts of law. Any disputes arising under or in connection with your use of this application or any information and material available on or through the application shall be subject to the exclusive jurisdiction of the courts of Oslo, Norway.
Reliability and accuracy:
Generative AI is essentially a statistical tool, trained on a limited set of information and not able to self-validate the truthfulness of its own answers. As such, any reliance on the AI Assistant’s responses is solely at the user’s own risk. DNV expressly disclaims any liability for decisions or damages arising from the use of AI Assistant’s responses to the fullest extent permitted by applicable law. Users should always consult the original source of information or a qualified DNV representative before making any decisions. It is therefore advised to always refer to the actual source of information or contact a DNV representative directly before taking decisions based on the responses of the Assistant. Our industry is complex and a proper understanding of the context motivating a user question can be essential. The AI Assistant may have limited capability in understanding this context. Owing to this, DNV cannot be liable for the accuracy of the answers provided by the AI Assistant.
Professional advice:
The application is trained on contents and information provided through DNVs own website, yet as a generative AI model, the AI Assistant will generate an individual response to each query with the contextual understanding it has gained at that point from the ensuing conversation. It is therefore important to note that, regardless of how confident the application appears in its responses, the AI Assistant does not provide professional advice from DNV in any capacity. The AI Assistant’s responses should not be interpreted as professional, legal, operational, or technical advice. For such matters, please consult with qualified DNV experts or relevant professionals directly. The AI Assistant is not able to give strategic, operational or technical support or advice, nor does it speak on behalf of DNV and should therefore not be regarded as the equivalent of a DNV employee or spokesperson. Should you require advice and support please get in touch with DNV experts directly here: https://www.dnv.com/maritime/contacts/options/
Impartiality:
DNV recognizes the importance of performing our classification, certification and verification services in an impartial and independent manner. DNV processes and practices are established with full commitment to complying with applicable impartiality requirements arising from laws, regulations, accreditations, and standards. However, due to the nature of generative AI, DNV cannot conclusively ensure that responses generated by the AI Assistant fulfill impartiality requirements in every instance. Users are therefore strongly advised to verify information from original sources and exercise discretion when acting on AI Assistant responses. DNV has established, and follows, internal policies and procedures that ensure that any actual or potential risk of conflict of interest is properly identified, mitigated, and documented. In the context of generative AI, users are therefore strongly advised to refer to the original source as DNV cannot conclusively ensure that responses generated by the AI assistant fulfill all impartiality requirements at any and all times.
For more details, please go here: https://www.dnv.com/about/public-impartiality-statement/
Personal data and tracking:
To improve your experience and enable recurring reviews and long-term improvements on the application, we store all queries received by users of the Assistant and responses provided by the Assistant and review these on a regular basis. These interactions are logged solely for the purpose of improving the service, and data will not be shared with third parties without explicit consent unless required by law.
As stated in the cookie policy, the application requires essential session cookies and operates with optional marketing cookies for tracking and later optimization of the user experience via Google Analytics.
In compliance with GDPR, the service does not collect or purposedly store personal identifiable information (PII). In addition, we request that users refrain from inputting any PII while interacting with the AI Assistant. We have implemented a detector in the application such that, in case PIIs are inadvertently inserted while conversing with the AI Assistant, the detector activates and the specific query containing said PIIs is not processed further. In this case users receive a standard answer from the AI Assistant reminding them not to insert PIIs. The detector is, however, itself based on Generative AI and as such affected by inherent limitations of the technology and susceptible to errors.
Data storage:
The DNV AI Assistant uses a jwt token (similar to a cookie) to represent the session a user is having with the application. This token is necessary to execute the functionality of having a conversation with the application. The conversation is stored on the Azure services within the DNV instance. Data hosting is solely in Europe. All conversations older than 18month will be deleted.
As the AI Assistant may generate responses that are incomplete, inaccurate, or not suited to your specific needs, users are strongly advised not to make any critical decisions based solely on its responses. Always verify the information with the original source or consult a qualified DNV professional for critical matters.