Higher Logic sees tremendous value in the use of artificial intelligence (AI) technology to deliver a superior experience to both our customers and their users of our community engagement and marketing automation products.
NOTE: Higher Logic is committed to the responsible use of AI in the delivery of these solutions in the following ways.
Data security and privacy
All Subscriber Data will continue to be used and processed in accordance with the applicable Master Service Agreement and Data Processing Addendum in place between Higher Logic and each customer. Higher Logic’s standard Data Processing Addendum is available here: Higher Logic Data Processing Addendum
Higher Logic continues its commitment to using Subscriber Data only for the purpose of providing services to that specific Subscriber (Higher Logic customer). Subscriber Data will not be accessible to other Higher Logic customers through the introduction of AI features.
When using an AI-enabled feature, the model will consider each user's existing permissions and look only at content that they have access to when generating a response.
Higher Logic's approach to AI
Subscriber Data will not be used to train or fine tune models without permission. If such permission is received, Higher Logic may use your Subscriber Data to train or fine tune models; however, the resulting models would be for your sole use and will not be used by any other Subscriber.
Higher Logic will require each customer to opt-in to each individual AI-enabled feature within the application.
Content that is generated by AI technologies will be identified as such and, where possible, references to the non-AI source content will be provided.
Functionality that enables human review and editing of the AI-generated content will be implemented where feasible.
All third-party AI models used by Higher Logic are stateless services. The services providing the models may temporarily retain interaction data for the evaluation of detected abuses; however, the data that is retained is not permanently stored by Higher Logic; nor is it used for further training or improving the AI models.
How will AI technology features provide value without using Subscriber Data to train or fine-tune a model?
Many of the use cases currently use AI technology to summarize or categorize content. The most popular AI models in use today are capable of summarizing small batches of content without further training or disclosure of personal data. When content is provided to the static models employed by Higher Logic, we only provide the minimal amount of content necessary to establish context.
How will Higher Logic limit the AI technology from making "hallucinations"?
"AI hallucination" is a situation in which the AI technology perceives a pattern and provides false statements in the output. Higher Logic intends to use Subscriber Data as context upon which the AI technology generates its content. By limiting the AI technology to a specific set of content, AI hallucination should be minimized. Additionally, Higher Logic will, where possible, provide the reference content that was used to generate content to allow for further validation of that generated content.
How will Higher Logic keep Subscriber Data from being exposed or used in public AI models?
When using pre-trained models, Higher Logic uses models that are deployed solely for the use of Higher Logic services. These models will not retain any of the contextual data used to respond to a query beyond the scope of that query interaction. The data that are exchanged with these models are not used to train the models.
Which Large Language Models is Higher Logic using?
Higher Logic has evaluated several large language models (LLMs) for use within its product. At present, Higher Logic plans to use Microsoft Azure AI services to privately deploy the models it will be using. Additional consideration is being given to the future use of privately deployed models provided by AWS (Amazon Web Services) Bedrock services.
In all cases, our usage will be consistent with:
Where can I get more information about Microsoft Azure as it relates to data privacy?
Microsoft provides additional details around data privacy as it relates to the use of OpenAI services within Azure in Data, privacy, and security for Azure OpenAI Service.