HUMAN and AWS Forge a New Trust Layer for the AI-Driven Internet

NEW YORK, NY – December 16, 2025 – In a move set to define the security standards for the next generation of the internet, cybersecurity firm HUMAN Security, Inc. today announced a pivotal collaboration with Amazon Web Services (AWS). The company's AgenticTrust platform will now provide cryptographic verification for AI agents built using the Amazon Bedrock AgentCore browser, establishing a foundational trust layer for the burgeoning world of automated, AI-driven web interactions.

This integration means that AI agents deployed on AWS infrastructure can be cryptographically signed, ensuring their actions are authenticated, policy-compliant, and secure. The initiative tackles a critical security vacuum created by the rapid emergence of autonomous AI agents that perform tasks ranging from complex research to executing commercial transactions.

“AI agents are redefining how businesses operate and interact across the web, creating powerful new opportunities, but also new trust and security requirements is paramount,” said Stu Solomon, CEO at HUMAN Security. “By using Amazon Bedrock AgentCore within AgenticTrust, we’re establishing a cryptographically verified framework that allows enterprises to securely build, deploy, and manage AI agents using Amazon Bedrock AgentCore. The result is a safer, more transparent ecosystem for legitimate automated interactions across the web.”

The Dawn of Agentic Commerce and Its Perils

The collaboration arrives at a critical inflection point in digital transformation: the rise of 'agentic commerce.' This new paradigm envisions a future where autonomous AI agents, not just humans, drive economic activity. These agents can anticipate consumer needs, research products, negotiate prices, and execute purchases, operating with a level of autonomy far beyond simple automation bots. The market for this technology is exploding, with some projections forecasting the global AI Agents Market to swell from over $40 billion in 2024 to nearly $250 billion by 2033.

However, this powerful new ecosystem introduces novel and sophisticated risks. Unlike predictable bots, advanced AI agents can learn, adapt, and make independent decisions. This creates vulnerabilities that traditional bot management and fraud prevention tools, designed for a different era, are ill-equipped to handle. The primary threats include advanced impersonation, where malicious agents masquerade as legitimate ones; the spread of misinformation; sophisticated fraud schemes; and the general misuse of automated systems at an unprecedented scale.

Without a robust method for verifying an agent's identity and authority, the digital marketplace could become a high-risk environment, eroding trust and hindering the adoption of what is otherwise a transformative technology. The partnership between HUMAN and AWS aims to directly address this challenge by building security into the very fabric of agentic interactions.

A Cryptographic Blueprint for Verifiable Agents

At the heart of the solution is the application of cryptographic verification to AI agent traffic. HUMAN's AgenticTrust platform leverages established standards, including public-key cryptography and HTTP Message Signatures (RFC 9421), to create a digital 'passport' for each agent originating from the Amazon Bedrock AgentCore browser. When an agent sends a request to a website or API, it includes a cryptographic signature that can be independently verified. This process confirms the agent's identity, its origin from a trusted platform like AWS, and its authorization to perform a specific action.

This verification provides a powerful layer of governance. Enterprises can now set granular, action-based controls for their agents. For example, a company could deploy a research agent with permissions to browse websites and gather information freely, but cryptographically block it from creating an account, filling out a form, or accessing a checkout page. This level of control is essential for businesses to safely embrace agentic commerce and turn AI-driven traffic into a trusted revenue channel.

Amazon Bedrock AgentCore provides the advanced, scalable platform for building and deploying these agents, with tools for memory, observability, and policy enforcement. The integration with AgenticTrust adds the crucial external verification layer needed for these agents to interact safely with the rest of the internet.

“At AWS, we’re committed to helping our customers deliver secure, seamless experiences powered by trusted automation,” said Kosti Vasilakakis, Principal Product Manager at AWS. “Through our collaboration with HUMAN, we’re making it easier for AI agents built with Amazon Bedrock AgentCore to operate responsibly across the web, reducing friction while giving businesses and developers confidence in every verified agent interaction."

From Theory to Practice: Secure Agents in the Wild

The need for this security framework is not theoretical. Major organizations are already deploying sophisticated multi-agent systems on AWS. The PGA TOUR, for instance, uses a system built on AgentCore to generate thousands of articles, increasing content writing speed by 1,000% and dramatically cutting costs. Similarly, Workday leverages an AgentCore-based tool to help users analyze complex financial data using natural language, saving hundreds of hours per month.

While these agents have already proven their value within controlled environments, the HUMAN Security collaboration unlocks their potential to interact safely with the broader digital ecosystem. An AI content agent can now be dispatched to browse the web for research with a verifiable identity, assuring websites that it is a legitimate, policy-abiding entity and not a malicious scraper bot. A financial analysis agent could, with proper permissions, interact with external data sources, with each request cryptographically signed and logged.

This assurance transforms agentic AI from a powerful internal tool into a safe and scalable external workforce. It allows businesses like retailers, financial institutions, and travel companies to deploy agents that can interact with partners, suppliers, and customers across the web without introducing unacceptable levels of risk.

Building a Foundation for Regulatory and Ethical AI

The implications of this technology extend beyond commerce and into the emerging landscape of AI governance and ethics. As regulators worldwide, including those behind the EU AI Act, begin to impose strict requirements on high-risk AI systems, the principles of accountability, transparency, and traceability are becoming legally mandated. The EU AI Act, for example, will require clear documentation, risk management, and human oversight for AI systems that can significantly impact individuals.

Cryptographically verifiable agents provide a technical foundation for meeting these demands. By creating an immutable, auditable log of an agent's identity and actions, the technology offers a concrete mechanism for accountability. If an agent makes an error or acts outside its prescribed policies, its actions can be traced back to their origin. This fosters a 'Know Your Agent' (KYA) environment, where enterprises can prove the identity and govern the behavior of their digital workforce.

This verifiable trust layer is not just about preventing fraud; it is about building a sustainable and ethical future for AI. As autonomous agents become more integrated into our daily lives and economies, the ability to trust their identity and actions will be as fundamental as the security protocols that underpin online banking and e-commerce today. This collaboration represents a critical step in building that essential infrastructure.