The integration of digital technologies across governance, commerce, and communication demands a more autonomous and resilient Internet, that protects against manipulation and surveillance. A Responsible Internet architecture must empower end users with fine-grained control over data handling, enabling a trust-aware, adaptive management of network behavior at the data plane. To this end, emerging technologies such as Programmable Data Planes, In-band Network Telemetry, In-Network Reinforcement Learning, and Intent-Based Networking are critical to achieving these goals. In this work, we present a Responsible Internet Proof-of-Concept that: Supports user-intent-driven path control and telemetry-based observability, integrates reinforcement learning agents within programmable forwarding devices for autonomous and secure data path selection, and demonstrates its feasibility and performance through deployment and evaluation on the international FABRIC testbed.