AI Search General QueriesIssue <!-- /*NS Branding Styles*/ --> .ns-kb-css-body-editor-container { p { font-size: 12pt; font-family: Lato; color: var(--now-color--text-primary, #000000); } span { font-size: 12pt; font-family: Lato; color: var(--now-color--text-primary, #000000); } h2 { font-size: 24pt; font-family: Lato; color: var(--now-color--text-primary, black); } h3 { font-size: 18pt; font-family: Lato; color: var(--now-color--text-primary, black); } h4 { font-size: 14pt; font-family: Lato; color: var(--now-color--text-primary, black); } a { font-size: 12pt; font-family: Lato; color: var(--now-color--link-primary, #00718F); } a:hover { font-size: 12pt; color: var(--now-color--link-primary, #024F69); } a:target { font-size: 12pt; color: var(--now-color--link-primary, #032D42); } a:visited { font-size: 12pt; color: var(--now-color--link-primary, #00718f); } ul { font-size: 12pt; font-family: Lato; } li { font-size: 12pt; font-family: Lato; } img { display: ; max-width: ; width: ; height: ; } } Queries from Customer 1. Please describe the purpose of the AI. Including what data and data subjects (i.e. clients, employees) are in scope.2. Is your AI system being trained, or was it developed, by using or processing personal data (including special categories of personal data)?3. Is the AI system a self-learning (autonomous) system?4. Is human oversight and control possible? (Fully, Partially or Not supervised by human operator?)5. How have the input/data-sets been developed to ensure diversity and representativeness of end-users and/or subjects in the data?6. How do you test and monitor for potential biases during the entire lifecycle of the AI system (e.g. biases due to possible limitations stemming from the composition of the used data sets (lack of diversity, non-representativeness).7. Could a low level of accuracy (within the algorithm and the data-sets) of the AI system result in critical, adversarial or damaging consequences?8. Who is responsible for:monitoring and assessing data quality & accuracy?monitoring and identifying bias?maintaining the algorithm?deciding when human control is required?reporting any errors following the above monitoring?9. Did you establish mechanisms that facilitate the AI system's auditability (e.g. traceability of the development process, the sourcing of training data and the logging of the AI system's processes, outcomes, positive and negative impact)?10. How are end-users or data subjects adequately made aware that a decision, content, advice or outcome is the result of AI, Machine Learning or Advanced Algorithm?Release<!-- /*NS Branding Styles*/ --> .ns-kb-css-body-editor-container { p { font-size: 12pt; font-family: Lato; color: var(--now-color--text-primary, #000000); } span { font-size: 12pt; font-family: Lato; color: var(--now-color--text-primary, #000000); } h2 { font-size: 24pt; font-family: Lato; color: var(--now-color--text-primary, black); } h3 { font-size: 18pt; font-family: Lato; color: var(--now-color--text-primary, black); } h4 { font-size: 14pt; font-family: Lato; color: var(--now-color--text-primary, black); } a { font-size: 12pt; font-family: Lato; color: var(--now-color--link-primary, #00718F); } a:hover { font-size: 12pt; color: var(--now-color--link-primary, #024F69); } a:target { font-size: 12pt; color: var(--now-color--link-primary, #032D42); } a:visited { font-size: 12pt; color: var(--now-color--link-primary, #00718f); } ul { font-size: 12pt; font-family: Lato; } li { font-size: 12pt; font-family: Lato; } img { display: ; max-width: ; width: ; height: ; } } AllResolution<!-- /*NS Branding Styles*/ --> .ns-kb-css-body-editor-container { p { font-size: 12pt; font-family: Lato; color: var(--now-color--text-primary, #000000); } span { font-size: 12pt; font-family: Lato; color: var(--now-color--text-primary, #000000); } h2 { font-size: 24pt; font-family: Lato; color: var(--now-color--text-primary, black); } h3 { font-size: 18pt; font-family: Lato; color: var(--now-color--text-primary, black); } h4 { font-size: 14pt; font-family: Lato; color: var(--now-color--text-primary, black); } a { font-size: 12pt; font-family: Lato; color: var(--now-color--link-primary, #00718F); } a:hover { font-size: 12pt; color: var(--now-color--link-primary, #024F69); } a:target { font-size: 12pt; color: var(--now-color--link-primary, #032D42); } a:visited { font-size: 12pt; color: var(--now-color--link-primary, #00718f); } ul { font-size: 12pt; font-family: Lato; } li { font-size: 12pt; font-family: Lato; } img { display: ; max-width: ; width: ; height: ; } } 1Q:- Please describe the purpose of the AI. Including what data and data subjects (i.e. clients, employees) are in scope. Ans:- Kindly review the documentation on "Machine learning relevancy in AI Search" to understand the purpose of AI https://docs.servicenow.com/bundle/tokyo-platform-administration/page/administer/ai-search/concept/machine-learning-relevancy-ais.html AI Search UX components record signals associated with user searches. These search signals include data on how search users interact with the search input field, auto-complete suggestions, facet and navigation tab filters, Genius Result answer cards, and search results that is used for machine learning relevancy tuning Search signal tables: https://docs.servicenow.com/bundle/tokyo-platform-administration/page/administer/search-administration/reference/search-signal-tables.html 2 Q:- Is your AI system being trained, or was it developed, by using or processing personal data (including special categories of personal data)? Ans:- No personal data is used in the signals for training machine learning relevancy in AI Search, as the signal data is aggregated https://docs.servicenow.com/bundle/tokyo-platform-administration/page/administer/search-administration/concept/search-signals.html 3 Q:- Is the AI system a self-learning (autonomous) system? Ans:- Yes 4.Q:- Is human oversight and control possible? (Fully, Partially or Not supervised by human operator?) Ans:- You cannot override the machine learning relevancy training, but you can change the relevancy scores in your Search Results using the following features to override them - Result improvement rules - Define rules with configurable trigger conditions to boost, block, or promote search results for specific searches. You can also boost search results for documents matching elements of the user context, such as a user's location or device. https://docs.servicenow.com/bundle/tokyo-platform-administration/page/administer/ai-search/concept/result-improvement-rules-ais.html Synonyms - Synonyms expand search queries to include additional terms with equivalent meaning or usage. Improve search recall by configuring synonym dictionaries and defining synonyms. https://docs.servicenow.com/bundle/tokyo-platform-administration/page/administer/ai-search/concept/synonyms-ais.html https://www.servicenow.com/community/ai-intelligence-articles/ai-search-synonyms/ta-p/2479589 5Q:- How have the input/data-sets been developed to ensure diversity and representativeness of end-users and/or subjects in the data? Ans:- The data used for machine learning relevancy is based on how all users of your instance interact with AI Search. 6Q:- How do you test and monitor for potential biases during the entire lifecycle of the AI system (e.g. biases due to possible limitations stemming from the composition of the used data sets (lack of diversity, non-representativeness). Ans:- The data used for machine learning relevancy is generated by all of your users when they interact with AI Search through the creation of search signals. 7Q:- Could a low level of accuracy (within the algorithm and the data-sets) of the AI system result in critical, adversarial or damaging consequences? Ans:- Each Search Profile has its own AIS Search Relevancy machine learning. Upon creating the Search Profile, it will use the default relevancy model, which provides good level of accuracy. In Vancouver we will also offer A/B Model Testing. We do compare relevancy models, to ensure the best performing relevancy model is used for the Search Profile. 8. Q:- Who is responsible for: monitoring and assessing data quality & accuracy? Ans:- The customer monitoring and identifying bias? Ans:-The customer maintaining the algorithm? Ans:- ServiceNow Team deciding when human control is required? Ans:- The customer reporting any errors following the above monitoring? Ans:- The customer 9Q:- Did you establish mechanisms that facilitate the AI system's auditability (e.g. traceability of the development process, the sourcing of training data and the logging of the AI system's processes, outcomes, positive and negative impact)? Ans:- These are all internal and confidential processes. The source of the training data is generated on your instances 10.Q:- How are end-users or data subjects adequately made aware that a decision, content, advice or outcome is the result of AI, Machine Learning or Advanced Algorithm? Ans:- Users are not made aware of the relevancy score, however this is shown in the AIS Preview tool that comes with the "Advanced AI Search Management Tools" application. The search results should improve over time whereby the information they are looking for is returned in the top 3 results. The content is displayed based on the documents that you have indexed. It uses the platform security to only return documents they can view, as per our documentation "Content security in AI Search" https://docs.servicenow.com/bundle/tokyo-platform-administration/page/administer/ai-search/concept/content-security-ais.html You can also created further sort option to display search results in an order determined by their field values. Applying a search result sort option overrides the default AI Search relevancy-based result order. You can define custom search result sort options for your AI Search applications. https://docs.servicenow.com/bundle/tokyo-platform-administration/page/administer/ai-search/concept/sort-options-srch-app-cfg-ais.html -- For other Ai search FAQ , you can refer below KB https://support.servicenow.com/kb?id=kb_article_view&sysparm_article=KB0959688