Artificial intelligence-generated podcasts open new doors to make science accessible: a mixed-method evaluation of quality and the potential for science communication

Area of Science:

• Cardiovascular Nursing
• Medical Education
• Science Communication

Background:

• Science podcasts are valuable for education and dissemination.
• Generative artificial intelligence (GenAI) enables rapid podcast creation.
• GenAI challenges like hallucinations can impact content trustworthiness.

Purpose of the Study:

• To explore the quality of AI-generated podcasts.
• To assess their potential for science communication.

Main Methods:

• Mixed-method evaluation of 10 AI-generated podcasts.
• Participant questionnaires and interviews.
• Participants were unaware of the AI origin during evaluation.

Main Results:

• Half of participants identified the AI nature of the podcasts.
• Strengths: understandable, engaging summaries.
• Weaknesses: perceived "American style" reducing credibility, inaccuracies, mispronunciations.
• Most appropriate for patients/public; potential for professionals with tailoring.

Conclusions:

• AI-generated podcasts can supplement scientific articles.
• They present a valuable alternative to traditional science podcasts.
• Recommendations: rigorous evaluation, transparency, referencing, and author acknowledgment.