Artificial Intelligence-Based Chatbots in Genetic Counseling Practice: Current Uptake, Utilization, and Perspectives

A survey of North American genetic counselors reveals that while AI chatbots are widely used for general purposes and viewed as promising tools for improving workflow efficiency and patient education, their clinical adoption remains low due to concerns regarding accuracy, patient comprehension, and a lack of structured training.

The Gist

Imagine genetic counseling as a high-stakes, deeply personal conversation between a guide (the genetic counselor) and a traveler (the patient) navigating a complex map of family history and DNA. Now, imagine trying to introduce a new tool: a robot assistant (an AI chatbot) to help the guide.

This paper is essentially a survey asking the guides: "How much do you use this robot? Do you trust it? And what do you think it can and can't do?"

Here is the breakdown of what the study found, using simple analogies:

1. The "Personal vs. Professional" Gap

The Finding: Most genetic counselors (about 76%) use AI chatbots in their personal lives, just like you might use them to write a grocery list or plan a vacation. However, very few (less than 9%) actually use or recommend them for clinical work with patients.

The Analogy: Think of it like a chef. Many chefs love using a fancy new electric knife at home to slice tomatoes for dinner. But when they are cooking for a restaurant, they stick to their trusted, sharp steel knives because they aren't sure the electric one is safe or precise enough for the customers.

2. Where the Robot is Allowed vs. Where it is Not

The Finding: When counselors do use these tools in the clinic, they mostly use them for "boring" or repetitive tasks, like gathering family history or explaining basic facts. They are very hesitant to let the robot handle the "heavy lifting," such as telling a patient they have a positive genetic result or explaining a confusing "Variant of Uncertain Significance" (a genetic finding where the answer isn't clear).

The Analogy: The counselors are willing to let the robot be the receptionist who hands out the brochure and takes down the address. But they absolutely do not want the robot to be the doctor delivering the diagnosis or holding the patient's hand during bad news. Only about 2.5% of counselors felt confident enough to let the robot deliver serious genetic news.

3. The "Trust but Verify" Problem

The Finding: The biggest worry for counselors is that the robot might not understand if the patient actually gets the information. They also worry the robot might give outdated or wrong information.

The Analogy: Imagine the robot is a tour guide in a museum. The counselors are worried the robot might say, "This painting is from 1920," when it's actually from 1910. Worse, the robot can't look the tourist in the eye and see the confused look on their face to know, "Oh, I need to explain that again differently." The counselors feel a human is needed to read the room.

4. The "Burnout" vs. "Safety" Tug-of-War

The Finding: Counselors are tired. They spend a lot of time on paperwork and repetitive tasks. They see the robot as a potential way to clear their schedule so they can focus on the patients who really need human connection. However, they are scared that if they rely too much on the robot, they might lose the "human element" that makes genetic counseling special.

The Analogy: It's like a firefighter who is exhausted from carrying heavy hoses. They want a machine to carry the hoses so they can focus on rescuing people. But they are terrified that if the machine breaks or malfunctions, the rescue will fail. They want the machine to help, but they don't want the machine to replace the firefighter.

5. The "Training Gap"

The Finding: Very few counselors (only about 8%) have received any formal training on how to use these AI tools. Most feel they are flying blind.

The Analogy: It's like giving a pilot a brand-new, high-tech cockpit with screens they've never seen before, but not giving them a manual or a simulator lesson. They know the plane can fly, but they don't know how to land it safely.

The Bottom Line

The paper concludes that while AI chatbots are like a promising new tool that could help genetic counselors work faster and reduce their stress, the profession isn't ready to hand over the keys yet.

The counselors say: "We are interested, but we need better training, clearer rules, and proof that the robot won't make mistakes before we let it talk to our patients." They see the robot as a helper, not a replacement.

Technical Summary

Technical Summary: Artificial Intelligence-Based Chatbots in Genetic Counseling Practice

Problem Statement
The field of genetic counseling is experiencing rapid growth, driven by increasing demand for genetic testing and personalized medicine. However, this surge has outpaced the availability of trained counselors, leading to high workloads, extended patient wait times, and increased provider burnout. Approximately 25% of a genetic counselor's time is consumed by administrative and repetitive tasks—such as pretest counseling, scheduling, and disclosing low-risk results—that do not strictly require specialized expertise. While digital tools like telehealth have offered some relief, the specific role of Artificial Intelligence (AI) chatbots in automating these tasks within genetic counseling remains underexplored. There is a critical need to understand the current adoption rates, confidence levels, and perceived utility of AI chatbots among genetic counselors to determine if they can effectively alleviate administrative burdens without compromising the "human element" of care.

Methodology
The study employed a cross-sectional survey design to evaluate the uptake and perspectives of AI-based chatbots in genetic counseling.

• Participants: The survey targeted board-certified/board-eligible genetic counselors ($n=1065$) and genetic counseling students ($n=310$) across North America. A total of 217 individuals completed the survey (160 counselors, 57 students), representing a response rate of 15.0% for counselors and 18.4% for students.
• Instrument: An online survey was developed de novo using Qualtrics, as no validated instruments existed for this specific context. The survey consisted of four sections: demographics, general chatbot usage, clinical genetics chatbot usage, and perceptions of benefits/limitations.
• Data Collection: Data was collected between October 2024 and April 2025.
• Analysis: Descriptive statistics, frequency analysis, and Chi-square tests of independence were used to analyze trends and associations. Open-ended responses were subjected to manual content analysis. Student responses were excluded from specific analyses regarding direct clinical experience.

Key Contributions
This study provides the first comprehensive assessment of the current landscape regarding AI chatbot integration in genetic counseling practice. Its primary contributions include:

1. Quantifying Adoption: Establishing the disparity between the high usage of general AI chatbots in non-clinical settings versus the low adoption of domain-specific clinical genetics chatbots.
2. Mapping Confidence: Identifying specific clinical tasks where counselors feel confident delegating to AI (e.g., family history gathering) versus those where trust is minimal (e.g., disclosing Variants of Uncertain Significance or positive results).
3. Identifying Barriers: Systematically cataloging the perceived limitations, including concerns over patient comprehension, data accuracy, and workflow integration.
4. Assessing Training Gaps: Highlighting the severe lack of formal AI education among current practitioners and trainees.

Results

• Usage Patterns: While 76.5% of participants used general AI chatbots outside of clinical practice (primarily for text composition and general inquiries), only 8.8% reported using or recommending clinical genetics chatbots in practice. Among those who did use clinical tools, the primary applications were communicating with at-risk family members (61.1%) and patient education (55.6%). The most utilized tool was the Genetic Information Assistant (Gia®) by Invitae.
• Confidence Levels: Confidence in chatbot capabilities varied significantly by task. The highest confidence was observed in gathering family history information (40.7% confident/highly confident) and serving as an education tool (38.2%). Conversely, confidence was extremely low for disclosing positive genetic results or Variants of Uncertain Significance (VUS), with only 2.5% of participants expressing confidence in chatbots for these tasks.
• Perceived Benefits vs. Limitations: The most significant perceived benefit was the reduction of repetitive tasks (84.6%), followed by allowing time for other tasks (72.3%) and decreasing provider burnout (50.8%). However, major concerns dominated the responses: 85.6% of participants were concerned about the chatbot's inability to assess patient comprehension, and 74.4% worried about the accuracy and up-to-date nature of the information provided.
• Training and Education: Only 8.2% of participants had received any training on clinical genetics chatbots. Despite this, there is a strong desire for structured education, with 80.0% requesting guidelines on ethical use and 74.4% requesting training on interpreting chatbot-generated data.
• Attitudes: While 81.0% of participants emphasized the critical importance of the "human element," 78.0% expressed some level of concern that AI could replace human counselors. Nevertheless, the majority (83.1%) expressed at least some interest in incorporating chatbots into their practice.

Significance and Claims
The paper claims that while AI chatbots hold promise as supplementary tools to enhance workflow efficiency and reduce administrative burdens, significant challenges must be addressed before widespread implementation can occur. The authors assert that the current low adoption rate is not due to outright opposition but rather stems from a lack of trust in the technology's accuracy, concerns regarding patient comprehension, and the absence of structured training and regulatory guidelines.

The study concludes that for chatbots to be successfully integrated, the field must move beyond general AI familiarity toward specific, validated clinical applications. The authors emphasize that genetic counselors should guide the involvement of chatbots to ensure these tools complement rather than compromise the quality of genetic counseling. They call for professional organizations to establish evidence-based recommendations and for educational programs to incorporate AI training to bridge the current knowledge gap. Ultimately, the research suggests that balancing technological efficiency with personalized, human-centric care is essential for the future of the profession.