Embedded Files
By bridging the gap between medical research and advanced machine learning, we build tools designed to capture the nuanced reasoning and standards required for high-impact editorial decisions.
WHO WE ARE
Rooted in Medicine. Driven by AI.
PRISM-AI is led by a multidisciplinary team of clinical professionals and artificial intelligence experts headquartered in London, Ontario. Our leadership combines over 15 years of experience in digital publishing and software commercialization with software engineers specializing in explainable AI and Natural Language Processing. As a Canadian-controlled private corporation, we prioritize data sovereignty and the ethical integration of technology into scholarly workflows. By bridging the gap between medical research and advanced machine learning, we build tools designed to capture the nuanced reasoning and standards required for high-impact editorial decisions.
WHAT WE DO
A Canadian-led standard for secure, ethical AI in biomedical publishing.
PRISM-AI is a Canadian-Controlled Private Corporation (CCPC) dedicated to data sovereignty and ethical innovation. Unlike platforms that process data internationally, 100% of our AI workloads and storage are housed exclusively in Canadian data centers to ensure full compliance with PIPEDA and global privacy standards. We are proud to be building the next generation of publishing tools right here in London, Ontario.
OUR LEADERSHIP
Meet Our Team
Our leadership combines over 15 years of experience in digital publishing and software commercialization with software engineers specializing in explainable AI and Natural Language Processing.
People Lead
June Doe
Audit
Jason Doe
Recruitment
Jonathan Doe
Risk
Jonathan Doe
Frequently Asked Questions
GENERAL OVERVIEW
What exactly is PRISM-AI?
PRISM-AI (Peer Review Insight and Screening Model) is an AI-powered tool designed to help medical and scientific journals manage the volume and complexity of new submissions.
It uses a GPT-4.1 language model that has been fine-tuned on authentic, anonymized peer review data to capture expert reasoning.
How accurate is the screening process?
Initial testing demonstrated an overall decision accuracy of 81.25%.
The model achieved 100% accuracy for manuscripts identified as "Reject," meaning it is highly effective at catching submissions deemed unsuitable by human reviewers.
AI-generated comments have shown a 0.82 cosine similarity to human feedback, outperforming standard models like ChatGPT.
EDITORIAL CONTROL
Does this replace human editors?
Absolutely not; PRISM-AI is built to complement, not replace, editorial judgment.
All AI recommendations are presented as suggestions, and the Journal Editor remains the final arbiter of scientific validity and all final decisions.
How does the system handle bias?
The algorithm evaluates manuscripts based solely on scientific merit and linguistic consistency, independent of commercial interests.
First Pass Research Reviewing Inc. periodically audits the system for systemic biases related to geography or institutional prestige.
Submissions from affiliates are flagged for manual human oversight to prevent favoritism.
DATA PRIVACY & SECURITY
How is our data protected?
The Company processes all information in strict accordance with the Personal Information Protection and Electronic Documents Act (PIPEDA) and Ontario privacy frameworks.
We employ industry-standard security technologies to protect information from unauthorized access.
Why are manuscripts retained for five years?
To preserve the integrity of the scientific record and prevent "journal shopping" of flawed data, reviewed manuscripts are retained for a five-year period.
This data is used for audit trails, retrospective accuracy analysis, and the identification of duplicate submissions.
CUSTOMIZATION & LOGISTICS
Can PRISM-AI be tailored to our journal’s specific voice?
Yes, the model is designed to adapt through fine-tuning on your journal's past editorial decisions and reviewer comments.
Through prompt engineering, editorial teams can provide specific instructions to ensure the AI's feedback aligns with the journal’s unique priorities and subject matter.
What are the governing laws for this service?
The service is owned by First Pass Research Reviewing Inc. and is governed by the laws of the Province of Ontario and the federal laws of Canada.
Are there any limitations I should be aware of?
The Services are provided on an "as-is" basis, and because they are powered by Large Language Models, they may occasionally produce "hallucinations" or factually incorrect outputs.
The system may yield False Positives (missing flaws) or False Negatives (incorrectly flagging quality work), which is why human oversight remains essential.
INSIGHT
Over the past 10 decades we've worked with every government across the globe to reduce emissions
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Proin malesuada pulvinar consectetur. Quisque malesuada eros sit amet aliquet tempus.
SPEAK TO US
Want to find out more about us?
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Proin malesuada pulvinar consectetur. Quisque malesuada eros sit amet aliquet tempus.
Report abuse