AIM Media House

FDA Launches Elsa 4.0 and HALO to Expand Agencywide AI Operations

FDA Launches Elsa 4.0 and HALO to Expand Agencywide AI Operations

FDA launched Elsa 4.0 and consolidated more than 40 internal systems into HALO as it expands AI-driven regulatory workflows.

The U.S. Food and Drug Administration (FDA) launched Elsa 4.0, the latest version of its internal artificial intelligence assistant, alongside a new unified data platform called HALO, according to a 05/06 announcement from the agency.

The platform consolidation combines more than 40 FDA application, submission, and operational systems into HALO, short for Harmonized AI and Lifecycle Operations for Data. FDA said it has begun integrating HALO directly with Elsa so employees can query internal systems and build workflows without manually uploading documents into chats.

The agency said the rollout is part of a broader modernization effort focused on regulatory operations, scientific review workflows, and internal productivity. FDA Commissioner Marty Makary, MD, MPH, said the new capabilities are intended to reduce administrative burdens for agency staff and allow them to focus more on scientific work.

Jeremy Walsh, Chief AI Officer at FDA, said Elsa would increasingly serve as the primary interface into FDA systems and data infrastructure. Walsh said FDA staff previously brought information into Elsa manually, but the system is now being positioned directly on top of agency data layers.

FDA Expands Internal AI Workflow Capabilities

FDA said Elsa 4.0 includes custom AI agents, document generation, quantitative data analysis, chart and graph creation, voice-to-text dictation, optical character recognition (OCR), and repository search across large internal document collections.

The agency also added a secure web search capability. FDA said the platform can access refreshed web data in responses while remaining isolated from the public internet.

According to FDA, Elsa operates inside a FedRAMP High Google Cloud Platform (GCP) environment and does not train on internal FDA data or information submitted by regulated companies. The agency said human reviewers remain involved throughout all AI-assisted workflows.

Elsa first launched in June 2025 as an internal productivity tool for FDA staff. FDA said the initial rollout was completed ahead of schedule and under budget. The original launch focused on scientific review support, summarization, and operational assistance.

FDA has also been testing AI-assisted clinical oversight workflows beyond internal administrative use. Earlier this month, the agency partnered with AstraZeneca and Amgen on a pilot program focused on real-time AI monitoring of clinical trials using cloud infrastructure and machine learning systems.

Healthcare and Pharma Companies Face Growing AI Pressure

The FDA rollout comes as pharmaceutical companies, clinical research organizations, and healthcare providers continue integrating AI into research, compliance, and operational systems.

Several industry executives and regulatory professionals responding to the agency’s LinkedIn announcement described HALO as a potential “single source of truth” environment for FDA data access and retrieval. Others pointed to the consolidation effort as a sign that AI-assisted workflows could become increasingly common in regulated healthcare operations.

Healthcare organizations have already been restructuring operational systems around embedded AI tools instead of standalone copilots. Hospitals and imaging providers expanded AI deployment across diagnostics, clinical review, and workflow management throughout 2025.

The FDA rollout also arrives as healthcare organizations face increasing scrutiny around AI governance, compliance standards, and oversight responsibilities. State governments and regulators have begun introducing new healthcare AI oversight frameworks as adoption accelerates across the sector.

Key Takeaways

  • Launch Elsa 4.0 to enhance FDA's internal AI-driven regulatory workflows.
  • Consolidate over 40 systems into HALO for streamlined data operations.
  • Integrate HALO with Elsa for improved staff efficiency and workflow automation.
  • Introduce custom AI features including document generation and data analysis.
  • Ensure secure web search capability while maintaining data isolation from the internet.