IBM patent application targets AI-controlled Additive Manufacturing
IBM, headquartered in Armonk, New York, USA, has published US patent application 2026/0177993 A1: ‘Managing Three-Dimensional Printers Using an Artificial Intelligence Enabled Decision Engine.’ The application outlines a system that uses sensor data and machine learning to compensate for changing internal and external conditions during the Additive Manufacturing process, thereby improving consistency and quality.
According to the application, the proposed system gathers data from a network of sensors monitoring factors such as ambient temperature, humidity, airflow and machine operating conditions. An AI-enabled decision engine analyses this information in real time and generates commands that adjust process parameters while a build is in progress.
The patent application states that the system could modify variables including build speed, deposition strategy and other machine settings to minimise defects caused by environmental fluctuations or equipment-related variations. It also describes the potential use of protective layers or other adaptive measures where required to maintain build quality.
IBM notes that conventional Additive Manufacturing machines can be affected by a range of variables, including nozzle position, material temperature, machine condition and surrounding environmental factors, all of which may influence dimensional accuracy, mechanical properties and overall build quality. The proposed AI-driven approach is intended to reduce these effects without requiring operator intervention.
The application also describes continuous learning capabilities, enabling the AI model to improve its recommendations through feedback gathered from previous builds and accumulated process data.
The patent application was filed on December 23, 2024, and published by the US Patent and Trademark Office on June 25, 2026.
