Analytics in Action News

Google’s Artificial Intelligence Voice Assistant Bests Siri, Alexa

Google Assistant has the most accurate artificial intelligence capabilities for providing medication information, with Siri and Alexa making swift improvements.

Google's artificial intelligence voice assistant bests Siri, Alexa

Source: Getty Images

By Jill McKeon

- A new study from Klick Applied Sciences suggests that people who use artificial intelligence-driven voice assistants for medication information will find the best results from Google Assistant. Meanwhile, Siri and Alexa are catching up, making significant comprehension improvements.

The 2021 study is a follow-up to a 2019 study by the same researchers, which first revealed that Google Assistant was the most accurate and effective tool when asked about the top 50 most dispensed medications in the US, both brand-name and generic.

“Interactive voice assistants are fulfilling a very legitimate need for people who are using them to help manage several aspects of their lives, including their health,” said Adam Palanica, PhD, co-author of the study, in a press release.

These improvements are especially noteworthy as the pandemic has driven up the value and use of digital health care services, along with advancements in artificial intelligence software. Accuracy for intelligent voice (or virtual) assistants (IVA) is especially important in helping patients who cannot type or read small text on medication labels, the study affirms.

In recent years, chatbots and virtual assistants have been noted as a potential entry point for integrating artificial intelligence into health care. In addition, virtual assistants and AI-powered chatbots have had a major impact on meeting patient needs virtually during the pandemic.

Researchers used the same audio recordings that they used to test the voice assistants in 2019, from 46 participants. All participants spoke English fluently. Accents did not play a significant role in accuracy, compared to the 2019 study where Alexa and Siri were eight to 11 percent less accurate for voices with foreign accents.

“Not only does this show that the AI systems have made tremendous progress in better detecting and understanding accents, but it tells us that the latest voice assistant technology is more usable for customers from a wider range of demographic groups, which is extremely important and good news given the diverse population,” said Palanica in the press release.

While the 2021 study still showed Google Assistant to be the most accurate voice assistant, both Siri and Alexa made significant improvements from 2019 to 2021, while Google’s accuracy remained almost the same. Siri made the most significant improvements, with a brand-name accuracy rate of 78.4 percent and generic accuracy at 75 percent, compared to 58.5 percent and 51.2 percent, respectively.

Like the 2019 study, Alexa lagged behind the other two voice assistants, with 64.2 percent brand-name accuracy and 66.7 percent generic accuracy, the study stated. However, these accuracy rates are significantly higher than the 2019 results, which showed 54.6 percent accuracy for brand-name medications and 45.5 percent for generic.

Google Assistant had 86 percent accuracy for brand-name medications, and 84.3 percent for generic medications, compared to 91.8 percent brand-name accuracy and 84.3 percent generic accuracy in 2019. The study pointed out that “no differences were found for generic names between 2019 and 2021, possibly indicating a ceiling effect in performance.” However, Google Assistant’s brand-name accuracy decreased by almost six percent.

“A voice assistant’s ability to accurately recognize proper medication names can have life-altering consequences for patients who use technology to help manage their health,” stated Yan Fossat, vice president of applied sciences at Klick Labs, and the study’s primary investigator.

“We continue to be encouraged by the progress of these findings and expect AI to grow more ‘intelligent’ as it gathers more data from users to update software algorithms. Currently, however, our research suggests that patients and consumers should not solely rely on voice platforms for accessing important medical information and advice.”