Why Amazon Hopes Generative AI Can Bring Alexa Back From Dead

Amazon's Echo/Alexa went from essential to irrelevant in hypertime. 

But Amazon is hoping that generative AI may be able to revive it - or something like it - by identifying new sources of revenue and profit based on user activity. The problem, natch, is that this is both intrusive and possibly illegal. But nice try. JL 

Scharon Harding reports in ars technica:

While voice assistants initially seemed to be a convenient, futuristic way to get information and perform basic tasks, they have barely graduated from that role. And the lack of evolution has left voice assistants surrounded by uncertainty. Google, for example, has shut down third-party Google Assistant smart displays and reportedly shifted Assistant manpower to Bard.Amazon has been trying to bring in revenue through Alexa partnerships. This newly reported use suggests Amazon is considering questionable techniques in the name of a unique generative AI experience.

While voice assistants initially seemed to be a convenient, futuristic way to get information and perform basic tasks, they have barely graduated from that role. And the lack of evolution has left voice assistants surrounded by uncertainty. Google, for example, has shut down third-party Google Assistant smart displays and reportedly shifted Assistant manpower to Bard. But while Google Assistant and Google's experimental Bard chatbot currently feel like different products with different uses, Amazon has dreams of uniting its generative AI efforts with its struggling Alexa business.

It's no secret that belts are tightening at Amazon, compounding interest in making Alexa a strong revenue source. Alexa was reportedly set to lose $10 billion in 2022, per an Insider report, and had failed to sufficiently engage users in ways that make Amazon money. Amazon is also enduring its largest round of layoffs and last week announced it is discontinuing Halo fitness and sleep trackers.

 

Can generative AI generate Alexa revenue?

Amazon reportedly tried incorporating more AI into Halo before killing it—like having trackers leverage a smartphone camera and computer vision to analyze and share user workout data with Amazon. We weren't eager to trust Amazon with such AI usage; however, Amazon is reportedly shifting some of that invasive AI energy to Alexa.

A report from Insider on Tuesday cited a "leaked document" titled "Alexa LLM [large language model] Entertainment Use Cases." It reportedly details plans to make Alexa more capable of "thinking vs. fetching from a database."

The AI, an Amazon spokesperson told Insider, isn't based on an open source model like versions being developed by other Big Tech companies but, rather, a proprietary LLM called Alexa Teacher Model. Alexa has already been using it for years, but Amazon is "building new models that are much larger and much more generalized and capable" to make Alexa "more proactive and conversational," according to Amazon's rep.

The internal document, Insider said, provides an example of what this beefed-up Alexa might be able to do. One sees Alexa creating a bedtime story using a prompt from a kid, like "cat and a moon." Amazon seems keen on using cameras to aid its AI, with Insider reporting that Amazon is exploring using an Echo Show smart display camera to identify a toy the child is holding and incorporate that into the story. This sort of intimate data collection for the use of voice assistant skills, however, would likely draw concern. Earlier this week, The Verge reported that Amazon workers "expressed pause" about incorporating computer vision into the Halo subscription service.

This storytelling feature could bring revenue by encouraging business partnerships. The leaked document reportedly mentions the hypothetical child holding a toy of Olaf from Frozen that Alexa would add to its story and named Lego "and others," per Insider, as potential partners.

Amazon has been trying to bring in revenue through Alexa partnerships with the likes of Domino's and Uber; however, this newly reported possible use suggests Amazon is considering questionable techniques (that would hopefully require user permission) in the name of a unique generative AI experience. The internal memo is said to discuss making the stories interactive by asking the user to make a corresponding illustration on an Echo Show display or to add to the AI tale.

The leaked memo also looks at ways to leverage Fire TVs to boost Alexa's helpfulness, Insider reported. Fire TV sets and streaming devices are reportedly selling well for Amazon and have the potential for revenue-driving experiences that are harder to replicate with, for example, smart speakers without screens. This includes letting users interact with digital content while connecting them to Prime Video and third-party streaming services.

Amazon has previously discussed developing Alexa to make better genre recommendations. And this all could help Amazon get detailed information on customer habits (Amazon says it doesn't sell customer data. But it could use insights for targeted marketing, for sharing non-user-specific trends with third parties, and to drive business decisions).

Insider said an internal Amazon memo discussed ways to use the Alexa LLM to "understand complex entertainment requests, context on content, and offer more accurate, personalized search results across video and audio."

An example cited is a user asking for content similar to Emily in Paris on Netflix but with less fashion, resulting in Alexa pointing to The Marvelous Mrs. Maisel, which happens to be a Prime Video original. If that Alexa user is on a device with a screen, the device could show an auto-playing trailer and "a visual map of what makes it a good recommendation. If the user follows up with 'more options with a female lead getting a dream job,' Alexa would give a list of shows based on the new request," Insider reported.

According to an example reportedly in the document, Alexa could also help users figure out the name of a show they're thinking of and, on Fire TV, recommend a free trial to a partner streaming service with the program, provide post-trial pricing upon request, and allow them to subscribe.

Other cited features by Insider include improved news and Q&A, with the document suggesting Alexa "present a summary of personalized news highlights for customers" and "help customers dig deeper into each story by retrieving related news and information." That related information could be, for example, a podcast Alexa recommends.

The clock is ticking

If there was ever a time for an Alexa revamp, it's now. We're not a fan of recording kids to boost a voice assistant's aptitude, though. Meanwhile, these possible uses aren't guaranteed to be used consistently by customers or drive sufficient revenue to keep the voice assistant afloat and allow for further development.

However, Amazon CEO Andy Jassy believes generative AI has recently advanced to a place with unprecedented possibilities.

During an Amazon Q3 earnings call on April 27, per a Seeking Alpha transcript, Jassy said:

These large language models and generative AI capabilities, they’ve been around for a while, but frankly, the models were not that compelling before about 6, 9 months ago. And they have gotten so much bigger and so much better much more quickly, that it really presents a remarkable opportunity to transform virtually every customer experience that exists and many that don’t exist that weren’t really that easily made possible before.

However, it will take a lot to make Alexa capable of doing tasks that are more advanced than simple commands, that users want to do frequently, and that bring revenue. Considering how shaky the voice assistant landscape looks, the clock is ticking for Amazon to figure all this out.