Company AI Risk Explorer
Loading the data…
Welcome to the Company AI Risk Explorer

This tool maps what large global companies say in public about AI risk. We reviewed public documents from roughly 200 of the world's largest companies. We coded every risk they acknowledge against the AI Risk Repository taxonomy.

Filter by company, sector, AI role, or risk. Compare what companies acknowledge against expert ratings of severity, sector vulnerability, and role responsibility.

← Back to the overview Company AI Risk Explorer

Companies × risks

Salesforce

Salesforce deploys AI at scale across its CRM products through Agentforce, Einstein, and Slack, integrating and orchestrating foundation models from OpenAI, Anthropic, and others. It also develops specialised models (CodeGen for code synthesis, XGen research LLMs) and has trained models on up to 1.5 trillion tokens.

We reviewed 13 public documents from Salesforce and found it acknowledges 13 of 24 AI risks, including discrimination and loss of privacy.

Sector: Information

Roles: AI deployer; Specialized AI developer; General-purpose AI developer

AI risks Salesforce acknowledges

  • Discrimination: “Safety: Mitigate bias… For example, an LLM can be guided to prevent the use of gender identity, age, race, sexual orientation, socioeconomic status, and other variables.” (How Salesforce Shapes Ethical AI Standards in the Agent Era)
  • Loss of privacy: “data masking is the process that replaces sensitive Personally Identifiable Information (PII) or proprietary business data with non-identifiable tokens before the prompt is sent to the LLM.” (Trusted AI: Key Principles | Salesforce)
  • Toxic content: “scan and categorize generated content in real time for signs of hate speech... harassment... Should harmful content be identified, the system's content filtering will automatically filter, block, or flag the output” (Trusted AI: Key Principles | Salesforce)
  • AI security vulnerabilities: “Data poisoning: Attackers can change how an AI responds by injecting malicious or misleading content into the training data or documents used in Retrieval-Augmented Generation (RAG).” (Salesforce Trusted AI Impact Report)
  • Capability & robustness: “These measures work to ensure that AI agents operate only within their intended domains, minimizing errors and reducing the likelihood of "hallucinations" or incorrect outputs.” (Salesforce Trusted AI Impact Report)
  • False information: “AI technologies, including generative AI, may create content that appears correct but is factually inaccurate or flawed...” (2025 Annual Report)
  • Overreliance & unsafe use: “Automation bias, where users overly trust AI outputs without critical analysis, risks amplifying errors and ethical concerns.” (Salesforce Trusted AI Impact Report)
  • Environmental harm: “AI-driven data center demand surge The rapid adoption of AI technology brings uncertainty to our data center emissions… the timing and scale of customer AI use… may significantly affect our emissions footprint.” (FY25 Stakeholder Impact Report)
  • Weapons & cyberattacks: “as AI technologies, including generative AI models, develop rapidly, threat actors are using these technologies to create new sophisticated attack methods that are increasingly automated, targeted and coordinated and more difficult to defend against.” (2025 Annual Report)
  • Loss of human agency: “Loss of human controls, a scenario where increasing autonomy in AI agents diminishes human oversight.” (Salesforce Trusted AI Impact Report)
  • Devaluation of human creativity: “may create content that... contains copyrighted or other protected material, and if our customers or others use this flawed or protected content... we may be exposed to... legal liability” (2025 Annual Report)
  • Disinformation & influence: “Salesforce evaluated the risks associated with facial recognition, including concerns about ... the potential misuse for intrusive surveillance.” (Salesforce Trusted AI Impact Report)
  • Governance failure: “Inadequate or ineffective AI development, deployment, content labeling or governance by us or others that result in controversy could also impair the acceptance of AI solutions or result in unintended performance of the services.” (2025 Annual Report)

Documents we reviewed