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

IBM

IBM provides enterprise AI infrastructure through its watsonx platform, offering tools, runtimes, and GPU-backed fine-tuning and hosting for customers. It also develops the Granite family of open-source foundation models (released under Apache 2.0), hosts third-party models (Meta Llama), and influences AI governance through founding roles in the Partnership on AI and the AI Alliance.

We reviewed 11 public documents from IBM and found it acknowledges 19 of 24 AI risks, including capability & robustness and AI security vulnerabilities.

Sector: Information

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

AI risks IBM acknowledges

  • Capability & robustness: “Poor model accuracy occurs when a model's performance is not sufficient to the task it was designed for... Low accuracy may occur if the model is not correctly engineered, or there are changes to the model's expected inputs.” (Foundation models: Opportunities, risks and mitigations)
  • AI security vulnerabilities: “Jail breaking: A jailbreaking attack attempts to break through the guardrails established in the model to perform restricted actions... Prompt injection attacks... produce unexpected output by manipulating the structure, instructions, or information contained in its prompt.” (Foundation models: Opportunities, risks and mitigations)
  • False information: “Hallucinations are the generation of factually inaccurate or ungrounded content with respect to the model's training data or input.” (Foundation models: Opportunities, risks and mitigations)
  • Loss of privacy: “ChatGPT suffered a bug and exposed titles and active users' chat history to other users... even more private data... first and last name, email address... last four digits of their credit card number” (Foundation models: Opportunities, risks and mitigations)
  • Discrimination: “Training an AI system on data with bias, such as historical or societal bias, could lead to biased or skewed outputs that can unfairly represent or otherwise discriminate against certain groups or individuals.” (Foundation models: Opportunities, risks and mitigations)
  • Transparency & interpretability: “Explanations for model output decisions might be difficult, imprecise, or not possible to obtain... Without clear explanations for model output, it is difficult for users... to understand and trust the model.” (Foundation models: Opportunities, risks and mitigations)
  • Toxic content: “Toxic output occurs when the model produces hateful, abusive, and profane (HAP) or obscene content... can adversely impact and harm people interacting with the model.” (Foundation models: Opportunities, risks and mitigations)
  • Devaluation of human creativity: “AI models to plagiarize existing work intentionally or unintentionally... claim the authorship or originality of works that were created by other people thereby engaging in plagiarism.” (Foundation models: Opportunities, risks and mitigations)
  • Overreliance & unsafe use: “Over-reliance occurs when a person puts too much trust in a model... accepting a model's output when the model's output is likely incorrect... over/under reliance can lead to poor decision making because of the misplaced trust in the AI system, with negative consequences.” (Foundation models: Opportunities, risks and mitigations)
  • Disinformation & influence: “The growth of generative AI has raised concern that this technology could disrupt major elections across the world. With AI deepfakes, a candidate's image can be smeared... Voters can be steered toward or away from candidates... a surge of AI deepfakes could erode the public's trust in what they see and hear.” (Foundation models: Opportunities, risks and mitigations)
  • Weapons & cyberattacks: “help in the planning and execution of a biological attack. Several test scenarios are mentioned such as using LLMs to identify biological agents and their relative chances of harm to human life.” (Foundation models: Opportunities, risks and mitigations)
  • Fraud & scams: “Generative AI models might be intentionally used to imitate people through deepfakes using video, images, audio, or other modalities without their consent.” (Foundation models: Opportunities, risks and mitigations)
  • Loss of human agency: “Impact on human agency: The autonomous nature of AI agents in performing tasks or taking actions could affect the individuals' ability to engage in critical thinking, make choices and act independently.” (AI agents: Opportunities, risks, and mitigations — IBM AI Ethics Board)
  • Inequality & unemployment: “Widespread adoption of foundation model-based AI systems might lead to people's job loss as their work is automated... Job loss might lead to a loss of income and thus might negatively impact the society and human welfare.” (Foundation models: Opportunities, risks and mitigations)
  • AI misalignment: “Alignment Misaligned Actions: AI agents can take actions that are not aligned with relevant human values, ethical considerations, guidelines, and policies.” (AI agents: Opportunities, risks, and mitigations — IBM AI Ethics Board)
  • Governance failure: “If the metrics do not measure the risk as intended, then the understanding of that risk will be incorrect and mitigations may not be appropriately applied. If the model's output is consequential, this may result in societal, reputational, or financial harm.” (Foundation models: Opportunities, risks and mitigations)
  • Unequal performance: “darker-skinned females are the most misclassified group (with error rates of up to 35% ). In comparison, the error rates for lighter-skinned were no more than 1%.” (Foundation models: Opportunities, risks and mitigations)
  • Environmental harm: “Impact on Environment: AI, and large generative models in particular, may produce increased carbon emissions and increase water usage for their training and operation.” (Foundation models: Opportunities, risks and mitigations)
  • Multi-agent risks: “background execution in multi-agent environments increases the risk of covert channels if input/output validation is weak.” (AI agents: Opportunities, risks, and mitigations — IBM AI Ethics Board)

Documents we reviewed