This is AI Frontiers. Today, integrity as a contract problem. On 2nd August 2026, the California AI Transparency Act becomes operative, the first U.S. mandate requiring providers of large generative AI systems to embed provenance disclosures in AI-generated content and offer a free public detection tool. Violations carry a $5,000 or per violation penalty, enforceable by the Attorney General. California's governor signed the enabling amendment, AB 8053, in October, and its start date was chosen deliberately to align with the EU AI Act's implementation timeline. That date matters less than what it confirms. AI has broken the old assumption that trusted systems produce trustworthy records. For years, firms treated integrity as the quiet leg of the security triad. Confidentiality got the headlines. Availability got the budget after every outage. Integrity sat in the background, a technical property of systems that, properly configured, kept data accurate enough to trust. That model held because integrity failures in conventional systems had a visible cause. A user changed a field, a script overwrote a file. Security teams could reconstruct the event path. Access controls, checksums, audit logs, all designed for alteration, not authorship. They can show that a record changed. They cannot establish that a newly generated output is accurate or grounded in a reliable source. AI breaks this at the root. Authorized systems can now generate, modify, and route information that looks legitimate while being wrong. The system is approved. The user is authorized. The workflow is legitimate. The record is still wrong. Three mechanisms drive this. Data poisoning corrupts training and retrieval pipelines. It rarely looks like sabotage. It looks like slightly biased rankings or distorted summaries. Drift inside approved workflows degrades model fit over time without any intrusion. And plausible but wrong output is the most expensive category because it passes human review more often than obvious nonsense. Corporate disclosure already reflects the shift. 72% of SP 500 companies disclosed AI as a material risk in their 2025 10Ks. Reputational risk, biased outcomes, unsafe outputs, was the most frequently cited concern, named by 38% of SP 500 firms. JP Morgan's 2026 10K names the failure mode outright. Inaccurate or biased output from rapid deployment and insufficient testing. The contract is now the firewall. Whoever controls the logs, the provenance trail, and the contractual definitions of acceptable error has a structural advantage. The supplier gets paid for speed. The customer often pays to verify. In AI markets, truth has become a billable input. That's AI Frontiers.