DigiCert Forecasts Tighter Controls on AI Agents

The company's annual security outlook predicts enterprises will demand verifiable identities and named owners for AI agents by 2027.

Updated on Nov. 21, 2026 in Cybersecurity

DigiCert Forecasts Tighter Controls on AI Agents

Live Poll

Do you trust large companies to keep their AI systems under human control?

DigiCert published its annual security outlook, containing 12 predictions, including that enterprises will require verifiable identities, named owners and explicit permissions for AI agents. The company, through Chief Technology Officer Jason Sabin, forecasts these controls as a condition of giving agents operational responsibility.

Why it matters

As organizations expand the use of AI agents, the forecast suggests they will need to show that agent actions were authorized, that someone is accountable, and that they can intervene if needed. These expectations could shape how suppliers and internal teams govern automated systems.

The forecast is drawn from DigiCert's annual outlook, which contains 12 predictions. The outlook does not provide adoption rates or baseline data for the AI agent controls it describes.

The players

DigiCert

DigiCert is a digital trust provider that publishes an annual security outlook covering identity, cryptography and content authenticity.

Jason Sabin

Jason Sabin is Chief Technology Officer at DigiCert, the company that published the security predictions.

The details

DigiCert forecasts wider use of short-lived credentials and continuous monitoring for AI agents, and expects agents to take on more testing, maintenance and lower-risk coding. The outlook also projects that large buyers will ask suppliers to inventory their cryptography and present a phased post-quantum migration plan, alongside wider use of signed content credentials under C2PA standards.

Timeline

  1. 2027: Businesses are expected to demand proof of AI agent controls.

The Tech Race

The forecast for signed content credentials under C2PA standards sits alongside Chrome's work on Merkle Tree Certificates, which is expected to prompt certificate providers and major websites to begin testing and early deployments. Together these point to a industry-wide push to build verifiable trust into digital identities and content.

Users and developers may see stricter requirements for how AI agents are identified, authorized and monitored in the software they use. Companies that supply enterprise customers may need to inventory their cryptography and present a phased post-quantum migration plan.

The takeaway

Organizations deploying AI agents should begin documenting who owns each agent, what it is permitted to do, and how it can be halted. Reviewing cryptography inventories now can also prepare teams for future supplier requirements.

Further reading

For related coverage of threats and defenses, see the Cybersecurity section.

Source note: This article includes information reported by Facebook.

Live Poll

Do you trust large companies to keep their AI systems under human control?