IBM's AI privacy explainer covers both individual and enterprise risks and the comparison is worth having

Most AI privacy conversations collapse individual and enterprise concerns into a single discussion that does not serve either audience well. IBM's AI privacy piece https://www.ibm.com/think/insights/ai-privacy covers both with enough specificity that you can identify which risks are most relevant to your situation.

The individual risks, sensitive data collection without meaningful consent, surveillance, bias and discrimination, personal data leakage from model outputs, are the consumer-facing concerns that drive most public discourse.

The enterprise risks, data exfiltration through AI tool inputs, IP leakage through code-generating tools, regulatory exposure from training on customer data, and the governance liability of deploying AI systems that make consequential decisions, are the commercial concerns that drive enterprise security posture.

The interesting overlap: employees using consumer AI tools at work without IT oversight create enterprise risks that look like individual privacy behaviour. A developer pasting client code into a public AI assistant for debugging is making an individual convenience choice with enterprise security consequences.

Most relevant privacy laws and best practices being summarised in the article makes it a useful reference rather than just a perspective piece.

In your experience, is the biggest AI privacy risk in your organisation coming from intentional policy violations, from tools being used without adequate governance, or from the tools themselves behaving in unexpected ways?