DeepScribe, a leader in AI-powered medical documentation solutions, announced today its new Trust and Safety Suite. The innovative features in DeepScribe’s latest addition to its ambient AI technology reflect the company’s initiative to set industry standards for AI in healthcare in safety, transparency, and ethics, and ensure that AI advancements are as trustworthy as they are revolutionary.
“As a company at the forefront of generative AI innovation in healthcare, DeepScribe recognizes the imperative of aligning technological advancements with core human values,” said Matthew Ko, co-founder, president and COO, DeepScribe. “These values shape our technology and business practices, ensuring that our innovations are aligned with the best interests of healthcare professionals and patients alike. We truly believe in AI for good and make sure we apply that every day across all we do.”
The Trust and Safety Suite includes three new pivotal features:
- Clinical Moments: Allowing users to trace AI-generated medical notes back to their origins in the clinical conversation, Clinical Moments ensures accuracy, transparency and provides healthcare professionals with a clear understanding of how the AI interprets and synthesizes information from the clinical dialogue. Clinicians can highlight a snippet anywhere in the transcript and DeepScribe automatically identifies the moment in the encounter in which that element of the note was discussed.
- Note Insights: Note Insights is based on real-world performance metrics, and is an audit dashboard that provides administrators with a snapshot of DeepScribe’s performance across an organization. This dashboard populates two key metrics: edits per 100 notes, which measures the text changes a clinician makes to a DeepScribe note, and DeepScribe Trust Score, a percentile ranking reflecting the trustworthiness and accuracy of notes based on the average edit distance compared to other users.
- Expert Human Audits: DeepScribe’s expert human audits team will become accessible to users and administrators – submitting generated notes to the Expert Human Audits team. This team will review notes and grade the outputs against DeepScribe’s clinical accuracy framework, and then provide customized suggestions to improve output accuracy.
Committed to Trust and Safety
“Since inception, DeepScribe has been unwaveringly dedicated to trust and safety, and this commitment extends beyond the introduction of our new Trust and Safety Suite,” stated Akilesh Bapu, DeepScribe CEO. “We have diligently implemented a comprehensive set of measures aimed at upholding the highest standards of safety and reliability in DeepScribe’s AI-powered solutions. Today’s announcement marks another significant stride in reaffirming our commitment to meeting and exceeding these standards.”
These measures include rigorous data privacy protocols achieved through SOC 2, continuous AI training and monitoring for accuracy and bias mitigation with the help of systems like AWS SageMaker and Bedrock, as well as ongoing collaboration with healthcare organizations like Covenant HealthCare to align DeepScribe technology with real-world clinical needs.
Today’s announcement lays the groundwork for future innovations and strategic initiatives across the company’s generative AI offerings that are responsible and ethical. DeepScribe is focused on enhancing system safety, increasing transparency, ensuring data privacy, and embedding human centric values in every aspect of its technology.