Snap IncSNAP

Snap Inc faces evolving cybersecurity threats including social-engineering attacks involving AI agents and agentic AI.

“In the ordinary course of business, we collect, store, use, and share personal data and other sensitive information, including proprietary and confidential business data, trade secrets, third-party sensitive information, and intellectual property (collectively, sensitive information). Our efforts to protect our sensitive information, including information that our users, advertisers, and partners have shared with us, may be unsuccessful due to the actions of third parties, including traditional “black hat” hackers, nation states, nation-state supported groups, organized criminal enterprises, hacktivists, and our personnel and contractors (through theft, misuse, or other risk). We and the third parties on which we rely are subject to a variety of evolving threats, including social-engineering attacks (for example by fraudulently inducing employees, contractors, vendors, AI agents, users, or advertisers to disclose information or facilitate access to our sensitive information or systems), malware, malicious code, hacking, credential stuffing, denial of service, and other threats. These attacks may use, or may be enhanced or facilitated by, artificial intelligence, including the use of image, video, or audio deepfakes, prompt injection or other manipulation of the inputs, instructions, or outputs of AI models, or agentic AI used to facilitate attacks or vulnerability discovery and exploitation at scale. While certain of these threats have occurred in the past, they have become m”

Snap Inc highlights risks from using AI to develop software including AI agents and AI-generated code.

“Moreover, our products and services, and the internal systems that support them and our business, rely on software, hardware, and other systems developed or maintained by our engineering teams and third parties (including open source software), and all of these have contained and will contain vulnerabilities, errors, bugs, or defects, which may or may not be detected by our teams or the respective third parties prior to our or their release, usage, or reliance on them. We also face risks from vulnerabilities, errors, bugs or defects through the use of AI to develop or maintain our software or systems and the use of AI within those software and systems, including AI agents and AI-generated code, as AI models may create flawed, incomplete, or inaccurate outputs, some of which may appear correct. This may happen if the inputs that the model relied on were inaccurate, incomplete, or flawed (including if a bad actor “poisons” the AI technology with bad inputs or logic), or if the logic of the AI technology is flawed (a so-called “hallucination”). Supply chain attacks have also increased in frequency and severity, and we cannot guarantee that third parties in our supply chain have not been compromised or that their systems, networks, or code are free from exploitable vulnerabilities, errors, bugs, or defects. In addition, rapid advances in AI have enabled the discovery of security vulnerabilities at increased speed and scale across our industry, which may outpace our or third parti”

Snap Inc discusses operational risks and dependencies from AI-enabled and agentic workflows in internal operations.

“bug, vulnerabilities, privacy incidents, security incidents, and AI-specific risks like prompt injection, all of which could result in litigation, regulatory investigations, or enforcements, and require us to expend significant resources, all of which could seriously harm our business. There is also no guarantee that these AI platforms and tools will provide the intended benefits to our business. We may also become operationally dependent on AI-enabled or agentic workflows used in our internal operations and product development. Because these tools may require access to large volumes of data and internal systems, they may take unintended actions, create unintended data movement or execution paths, or otherwise disrupt important workflows, including in ways that are difficult to predict, monitor, or remediate. In addition, if third-party AI providers change pricing, functionality, availability, security posture, or contractual terms, or if workflows tied to individual personnel access are not effectively transitioned, we may incur significant costs and business disruption in migrating, replacing, or decommissioning related workflows.”

Snap warns that evolving threats including agentic AI are used to facilitate attacks on sensitive information.

“In the ordinary course of business, we collect, store, use, and share personal data and other sensitive information, including proprietary and confidential business data, trade secrets, third-party sensitive information, and intellectual property (collectively, sensitive information). Our efforts to protect our sensitive information, including information that our users, advertisers, and partners have shared with us, may be unsuccessful due to the actions of third parties, including traditional “black hat” hackers, nation states, nation-state supported groups, organized criminal enterprises, hacktivists, and our personnel and contractors (through theft, misuse, or other risk). We and the third parties on which we rely are subject to a variety of evolving threats, including social-engineering attacks (for example by fraudulently inducing employees, users, or advertisers to disclose information or facilitate access to our sensitive information or systems), malware, malicious code, hacking, credential stuffing, denial of service, and other threats. These attacks may use, or may be enhanced or facilitated by, artificial intelligence, including the use of image, video, or audio deepfakes, or agentic AI used to facilitate attacks or vulnerability discovery and exploitation at scale. While certain of these threats have occurred in the past, they have become more prevalent and sophisticated in our industry, and may occur in the future. Because of our prominence and value of our sensi”

Snap uses AI agents to generate code, answer support questions, find bugs, and improve efficiency.

“5 Transforming How We Operate to Drive Efficiency We are refining our operating model using a jobs to be done framework, mapping critical work to a combination of highly focused teams and increasingly capable AI agents. Focused teams aligned to top strategic priorities are allowing us to move faster, execute more efficiently, and better serve our community, partners, and advertisers. 1M+ 65%+ 7,500+ questions/month answered by support of new code generated by AI bugs found by code review agent agent Level the playing field Focus on what matters Execute faster AI gives small, creative teams the Free up teams to concentrate on Increase velocity of product ability to do work that once higher-value work and development — as demonstrated required much larger organizations. highest-priority strategic initiatives. by rapid progress on Snapchat+ and Snap Lite infrastructure improvements.”

Snap notes AI including agentic AI can enhance or facilitate cybersecurity threats and attacks.

“In the ordinary course of business, we collect, store, use, and share personal data and other sensitive information, including proprietary and confidential business data, trade secrets, third-party sensitive information, and intellectual property (collectively, sensitive information). Our efforts to protect our sensitive information, including information that our users, advertisers, and partners have shared with us, may be unsuccessful due to the actions of third parties, including traditional “black hat” hackers, nation states, nation-state supported groups, organized criminal enterprises, hacktivists, and our personnel and contractors (through theft, misuse, or other risk). We and the third parties on which we rely are subject to a variety of evolving threats, including social-engineering attacks (for example by fraudulently inducing employees, users, or advertisers to disclose information or facilitate access to our sensitive information or systems), malware, malicious code, hacking, credential stuffing, denial of service, and other threats. These attacks may use, or may be enhanced or facilitated by, artificial intelligence, including the use of image, video, or audio deepfakes, or agentic AI used to facilitate attacks or vulnerability discovery and exploitation at scale. While certain of these threats have occurred in the past, they have become more prevalent and sophisticated in our industry, and may occur in the future. Because of our prominence and value of our sensi”

Snap warns of evolving threats including agentic AI used to facilitate attacks.

“In the ordinary course of business, we collect, store, use, and share personal data and other sensitive information, including proprietary and confidential business data, trade secrets, third-party sensitive information, and intellectual property (collectively, sensitive information). Our efforts to protect our sensitive information, including information that our users, advertisers, and partners have shared with us, may be unsuccessful due to the actions of third parties, including traditional “black hat” hackers, nation states, nation-state supported groups, organized criminal enterprises, hacktivists, and our personnel and contractors (through theft, misuse, or other risk). We and the third parties on which we rely are subject to a variety of evolving threats, including social-engineering attacks (for example by fraudulently inducing employees, users, or advertisers to disclose information or facilitate access to our sensitive information or systems, malware, malicious code, hacking, credential stuffing, denial of service, and other threats. These attacks may use, or may be enhanced or facilitated by, artificial intelligence, including the use of image, video, or audio deepfakes, or agentic AI used to facilitate attacks or vulnerability discovery and exploitation at scale. While certain of these threats have occurred in the past, they have become more prevalent and sophisticated in our industry, and may occur in the future. Because of our prominence and value of our sensit”