Unit of Urology, Department of Neurosciences, Reproductive Sciences, and Odontostomatology, University of Naples "Federico II", 80131 Naples, Italy.
Unit of Urology, Department of Woman, Child and General and Specialized Surgery, University of Campania "Luigi Vanvitelli", 80131 Naples, Italy.
Int J Environ Res Public Health. 2023 Feb 1;20(3):2612. doi: 10.3390/ijerph20032612.
The aim of this study was to evaluate the ability of Google Assistant, Alexa, and Siri to recognize and answer questions about male sexual health. Each VA was tested on a smart speaker: Alexa on Amazon Echo Dot 4th Gen., Google Assistant on Google Home Mini, and Siri on Apple HomePod. A pool of patients' frequently asked questions regarding erectile dysfunction (ED), premature ejaculation (PE), Peyronie's disease (PD), male infertility, and other aspects of male sexual health were identified by authors. The recognition of question was evaluated ("yes" or "not"). For each recognized question, the response characteristics (domains) were rated on a scale from 0 to 10 (according to the quality). We chose the recognition rate of the questions as the primary outcome and the quality of the answers as the secondary outcome. Overall, the best VA in recognizing questions was Siri, with a total of 83.3% questions compared with 64.0% for Alexa ( = 0.024) and 74.0% for Google Assistant ( = 0.061). Siri was associated with a significantly higher recognition rate than Alexa for PE (80% vs. 40%; = 0.002) and PD (66.7% vs. 33.3%; = 0.010). The quality of the responses was classified as low in 57 out of 105 cases (54.3%), intermediate in 46 cases (43.8%), and high in only 2 cases (1.9%), highlighting an overall intermediate-low quality of the answers. Male infertility was the condition associated with the highest mean scores in "Targeted response to the problem" (7.32 ± 2.57), "Scientific correctness of the answer", (5.9 ± 2.76) "Completeness of the answer" (5.14 ± 2.56), and "Understandability of the response for a patient" (5.3 ± 2.51) domains. Siri was associated with significantly higher scores than Alexa ( < 0.05) in several domains of all conditions evaluated. The question recognition rate of VAs is quite high; however, the quality of the answers is still intermediate-low. Siri seems superior to Alexa in both question recognition and response quality. Male infertility appears to be the sexual dysfunction best addressed by VAs.
本研究旨在评估 Google Assistant、Alexa 和 Siri 识别和回答男性健康问题的能力。每个 VAs 在智能音箱上进行测试：Alexa 在亚马逊 Echo Dot 4 代上，Google Assistant 在 Google Home Mini 上，Siri 在 Apple HomePod 上。作者通过患者经常询问的问题，确定了勃起功能障碍 (ED)、早泄 (PE)、佩罗尼氏病 (PD)、男性不育和其他男性健康方面的问题。评估了问题的识别情况（“是”或“否”）。对于每个识别出的问题，根据质量从 0 到 10 对回答的特征（领域）进行评分。我们选择问题的识别率作为主要结果，答案的质量作为次要结果。总体而言，在识别问题方面，表现最好的 VAs 是 Siri，总共识别了 83.3%的问题，而 Alexa 为 64.0%（ = 0.024），Google Assistant 为 74.0%（ = 0.061）。Siri 在识别 PE（80%比 40%； = 0.002）和 PD（66.7%比 33.3%； = 0.010）方面的识别率明显高于 Alexa。在 105 个案例中，有 57 个（54.3%）被归类为低质量，46 个（43.8%）为中质量，只有 2 个（1.9%）为高质量，这突出表明答案的整体质量中等偏低。男性不育是与“针对问题的针对性反应”（7.32 ± 2.57）、“答案的科学正确性”（5.9 ± 2.76）、“答案的完整性”（5.14 ± 2.56）和“患者对答案的理解”（5.3 ± 2.51）等领域的平均分数最高的疾病。在评估的所有情况下，Siri 在多个领域的得分均明显高于 Alexa（<0.05）。VAs 的问题识别率相当高；然而，答案的质量仍然中等偏低。Siri 在问题识别和响应质量方面似乎都优于 Alexa。男性不育似乎是 VAs 最能解决的性功能障碍。
