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

Amazon

Amazon provides AI infrastructure through AWS (cloud compute, Trainium/Inferentia accelerators, and managed GenAI services including Bedrock). It develops general-purpose foundation models (Nova/Titan) and deploys third-party models at scale through Bedrock, while also building specialised AI for healthcare NLP and warehouse robotics.

We reviewed 10 public documents from Amazon and found it acknowledges 10 of 24 AI risks, including AI security vulnerabilities and toxic content.

Sector: Trade, Transportation, and Utilities

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

AI risks Amazon acknowledges

  • AI security vulnerabilities: “LLM-PIEval, a benchmark containing indirect prompt-injection attacks for LLMs... Attacks targeting sensitive APIs (e.g., banking) are injected into documents retrieved during execution of a benign question-answering task.” (Amazon Nova and our commitment to responsible AI)
  • Toxic content: “prohibits the use... (4) for harm or abuse of a minor, including grooming and child sexual exploitation” (Amazon 2025 Proxy Statement)
  • Dangerous capabilities: “Automated AI R&D could also accelerate the development of models that pose enhanced CBRN, Offensive Cybersecurity, or other severe risks... fully automating the research, development, and deployment of frontier models that will pose severe risk such as accelerating the development of enhanced CBRN weapons and offensive cybersecurity methods.” (Amazon's Frontier Model Safety Framework)
  • Weapons & cyberattacks: “risks that would arise from the use of a model by malicious actors to compromise digital systems with the intent to cause harm... enable a moderately skilled actor... to discover new, high-value vulnerabilities and automate the development and exploitation of such vulnerabilities.” (Amazon's Frontier Model Safety Framework)
  • Discrimination: “a dataset... prompts that test for bias across profession, gender, race, religion, and political ideology... In 2023, we introduced TANGO... tests for misgendering of transgender and nonbinary (TGNB) people” (Amazon Nova and our commitment to responsible AI)
  • Loss of privacy: “Our set of risk categories includes the following: sensitive-data exfiltration” (Amazon Nova and our commitment to responsible AI)
  • Capability & robustness: “detect hallucinations in model responses that are not grounded in enterprise data or are irrelevant to the users' queries” (Amazon 2025 Proxy Statement)
  • Disinformation & influence: “prohibits the use of AWS AI services... (1) for intentional disinformation or deception” (Amazon 2025 Proxy Statement)
  • False information: “AWS also applies AR to GenAI itself in order to help manage the problem of hallucinations.” (Amazon's Frontier Model Safety Framework)
  • Devaluation of human creativity: “our and our customers' development and use of artificial intelligence may result in increased claims of infringement or other claims, including those based on unauthorized use of third-party technology or content.” (Amazon 2024 Annual Report)

Documents we reviewed