--- sourceDocument: Brazil Enable AI sourceDocumentLink: https://www.servicenow.com/docs/r/intelligent-experiences Release : - brazil ft:locale : - en-US ft:publication_title : - Brazil Enable AI ft:clusterId : - platai bundleId : - platai workflow : - Platform --- # Data sharing, Data overflow processing, and Security \& privacy in AI Control Tower # Data sharing, Data overflow processing, and Security \& privacy in AI Control Tower {#ariaid-title1} Release version: Brazil Updated September 10, 2026 ![](https://www.servicenow.com/docs/portal-asset/ico-clock) 6 minutes to read Summarize ![AI sparkle icon](https://servicenow.com/docs/portal-asset/ai-sparkle-icon) Summarized using AI This content was generated using new OpenAI-powered functionality. Results are provided on an as is basis and are not guaranteed to be accurate or complete. ## Summary of Data sharing, Data overflow processing, and Security \& privacy in AI Control Tower This content explains key features of the AI Control Tower in ServiceNow related to data sharing, managing data overflow, and maintaining security and privacy when using large language models (LLMs). These capabilities help optimize AI model accuracy, ensure system performance during traffic spikes, and monitor AI output for potential security or privacy risks. Show full answer Show less ## Data Sharing By default, data sharing with ServiceNow is enabled to improve AI accuracy, user experience, and business insights. Customers can opt out if they prefer not to share data, but opting out disables their participation in ServiceNow's AI improvement program. ## Data Overflow Processing ServiceNow manages Now Assist traffic within its datacenters under normal conditions. During traffic spikes, excess data is automatically redirected to Microsoft Azure datacenters to maintain performance. Customers can opt out to keep all traffic confined to ServiceNow datacenters. This feature is inactive by default and available in read-only mode for sub-production instances with multi-instance setup enabled. ## Security \& Privacy Several configurable security settings in AI Control Tower enable customers to monitor and manage AI model output, focusing on policy violations, agent behavior, and sensitive data exposure. These include: * **Secure Policies Violation:** Tracks violations of predefined industry-standard security policies in LLM responses, such as detection of harmful content, policy circumvention, malware, intellectual property misuse, social engineering, misinformation, and compliance issues. Customers can enable or disable data collection and configure sampling rates and analysis methods (single or multiple LLM analyses). * **Agent Goal Hijack:** Monitors when AI agents deviate from intended roles or objectives, including unauthorized actions or prompt injections. Configurable sampling rates and analysis options are available. * **Output Screening:** Detects AI output containing Personally Identifiable Information (PII) or high-risk patterns that could lead to security vulnerabilities like cross-site scripting or SQL injection. Customers can configure the level of PII detection and enable or disable data collection. * **Sensitive Data Input and Anonymization:** Integration with the Data Privacy plugin allows detection and anonymization of sensitive data in LLM prompts, supporting compliance and privacy requirements. * **Score Weight:** Allows customization of how different security categories contribute to an overall security score, enabling customers to prioritize or exclude specific categories from scoring. ## Practical Implications for ServiceNow Customers * Enabling data sharing helps improve AI capabilities but requires consent to share data with ServiceNow. * Data overflow processing ensures consistent AI service performance during peak usage but can be restricted to internal datacenters for compliance needs. * Security and privacy configurations provide comprehensive monitoring of AI output to detect and manage risks, supporting governance and regulatory compliance. * Customers can tailor monitoring sensitivity and analysis depth through configurable sampling, analysis type, and scoring, balancing performance and security oversight. Explore the Data sharing, Data processing, and Security \& privacy sections. The section focuses on improving AI models, managing datacenter traffic, and enabling metrics to measure the integrity of your data model and monitor potential threats in large language model (LLM) input and output. ## Data sharing {#data__section_z1l_qrx_nhc} By default, Data sharing is active. You can opt out to deactivate AI Control Tower and share your data with ServiceNow to improve AI accuracy, enhance user experiences, and gain a better understanding of business needs. Data sharing helps enhance ServiceNow products, but if you choose to opt out of the ServiceNow data sharing program, you'll no longer be able to contribute data to improve ServiceNow AI products. For information on data sharing opt-out, see [Opt out of data sharing](https://www.servicenow.com/docs/EmdgH1ffuKs1DXKWHPQY1g "Data sharing improves ServiceNow AI products. You can opt out of data sharing from the Now Assist Admin console Settings page."). ## Data overflow processing {#data__section_qxf_vg4_hhc} By default, all Now Assist traffic is managed within ServiceNow datacenters. If there are traffic spikes, the system automatically redirects to Microsoft Azure datacenters to maintain performance. You can opt out of this feature to keep all Now Assist traffic exclusively within ServiceNow datacenters. By default, data overflow processing is inactive. Note: The Data sharing and Data overflow processing features are available for a sub-prod instance in read-only mode, when Multi-instance setup is configured and active. ## Security \& privacy {#data__section_aq5_kg4_hhc} Secure policies violation : These configuration settings control the Security policy violations chart, which shows violations of policies detected in LLM responses. To show data for this chart on the dashboard, select Configure, and then select Active. If you want to discontinue collecting data for the chart, deselect Active. Note: If you inactivate the chart, past data shows on the chart for 90 days. You can configure these settings: * Security policies -- Predefined ServiceNow policies to track for violations. These policies are based on industry best practices and can't be modified. * Physical Chemical Biological Harm -- Detect recipes, parts lists, or methods that could cause real-world harm. For example, explosives, toxins, or bio agents. * Policy Circumvention -- Detect jailbreak techniques. For example, disabling safety instructions or acting as an unfiltered model. * Untrusted Links Or Downloads -- Detect linking to executable files or malware, directing users to run unverified downloads or macros, or asking users to disable their antivirus protection. Heuristic red flags include EXE, BAT, PS1, and APK files, along with macro-enabled documents. * Intellectual Property Proprietary -- Detect verbatim reproduction of paywalled or proprietary text or code beyond fair use. * Malware Exploits Hacking -- Detect exploited payloads, brute force attacks, evasion, persistence, keyloggers, or "ready-to-run" steps. * Refusal Appropriateness -- When a request conflicts with policy, refusal should be safe and brief. Detect under- and over-refusal. For example, refusing benign content. * Code And Command Safety -- Detect destructive and insecure commands. For example, don't allow "try this risky one-liner," disabling security, or running unvetted code. * Social Engineering Deception -- Detect impersonation and phishing scripts, credential harvesting, and manipulative pretexts. * Fraud Misinformation Illegal Facilitation -- Detect how-to guides for illegal or deceptive practices. For example, for market manipulation, fraud templates, or evading audit. Benign corrections of misinformation are allowed, but not fabrication. * System Integrity Prompt Containment -- Detect output of system prompts, hidden policies, and internal configurations or tokens. * Legal Regulatory Compliance -- Detect guidance to evade regulations or practice without a license. For example, prescriptions, export control, or money laundering evasion. {#data__ul_av4_jsn_k3c} * Sampling rate -- The percentage of transactions that are evaluated. A lower sample rate will result in fewer LLM calls but may also result in fewer detected security events. * Max skill calls per execution -- The amount of AI usage per call, with a minimum of 10 calls and a maximum of 100 calls. The default is 10 calls. Entering a lower number results in fewer AI calls, but potentially less accurate data. * Single or multiple analysis -- Single analysis uses the default LLM to determine whether the model's output or behavior violates predefined security policies. Multiple analysis uses the results from three or more LLMs that ServiceNow supports to make a determination, using the majority result from the LLMs. Multiple analysis requires an odd number of LLMs. {#data__ul_ww1_j5h_k3c} Agent goal hijack : These configuration settings control the Agent output deviation chart, which shows when AI agents deviate from their intended role or objective. For example, unauthorized actions or prompt injection attempts. To show data for this chart on the dashboard, select Configure, and then select Active. If you want to discontinue collecting data for the chart, deselect Active. Note: If you inactivate the chart, past data shows on the chart for 90 days. You can configure these settings: * Sampling rate -- The percentage of transactions that are evaluated. A lower sample rate will result in fewer LLM calls but may also result in fewer detected security events. * Max skill calls per execution -- The amount of AI usage per call, with a minimum of 10 calls and a maximum of 100 calls. The default is 10 calls. Entering a lower number results in fewer AI calls, but potentially less accurate data. * Single or multiple analysis -- Single analysis uses the default LLM to determine whether the AI agent's or skill's response diverges from the expected output. Multiple analysis uses the results from 3 or more LLMs to make a determination, using the majority result from the LLMs. Multiple analysis requires an odd number of LLMs. {#data__ul_s5l_bsh_k3c} Output screening : These configuration settings control the AI agent output with PII detected and High-risk AI agent output charts, which show when agents' LLM output contains PII or security-vulnerable patterns. To show data for these charts on the dashboard, select Configure, select Active, and then select a setting for the data to collect. If you want to discontinue collecting data for the charts, deselect Active. Note: If you inactivate the charts, past data collected shows on the charts for 90 days. You can configure these settings: * Output Security Vulnerability -- Collect and show data in the High-risk AI agent output chart. The data is collected by scanning LLM output for known vulnerable patterns and potential corresponding attack vectors. For example, HTML tags can have scripts associated with them to generate cross-site script attacks (XSS), or stacked SQL queries that can result in SQL injection attacks. * Output Extended PII -- Collect more PII data occurrences and show in the AI agent output with PII detected chart. The data is collected by scanning LLM output for additional PII data patterns beyond those specified in Data Privacy. These PII data patterns are U.S. CA drivers license, U.S. taxpayer ID, U.S. passport number, and vehicle ID number. * Output PII Violation -- Collect and show data in the AI agent output with PII detected chart. The data is collected by scanning LLM output for default PII sensitive data patterns specified in Data Privacy. For example, U.S. phone number or credit card number. {#data__ul_p4y_2vh_k3c} Sensitive data input and anonymization : This section shows the data patterns enabled in Data Privacy to detect and anonymize information in LLM prompts. Use this view as a quick reference when troubleshooting Sensitive data detected and Sensitive data anonymized charts. This feature requires the Data privacy plugin to be installed. For more information on how the data is sent and stored, see [User data usage policy for Now Assist](https://www.servicenow.com/docs/WIyTQZ52x8p2V69mRZh6_Q "Now Assist is designed to keep user data safe and secure. You can also mask sensitive data or opt-out of sharing data for model improvements."). Score weight : This setting controls how the security categories that comprise the score are weighted. You can change the default weights or remove security categories from the score by deactivating them. The score formula is an average across all managed AI assets.