Amazon plans to "upgrade" Alexa to join the booming generative AI battle

According to sources familiar with the matter, Amazon (AMZN.US) is using artificial intelligence to upgrade its 10-year-old Alexa voice assistant and plans to charge a monthly subscription fee to offset the technology costs.

Amazon is set to launch a more conversational version of Alexa later this year, as revealed by two insiders. This move is aimed at better competing with new generative AI chatbots from companies like Google and OpenAI. One source mentioned that Alexa services will not be included in the annual $139 Prime subscription, and Amazon has not yet determined the pricing.

Amazon declined to comment on its plans for Alexa.

Ten years ago, Amazon's introduction of the Alexa voice assistant in 2014 amazed consumers, but in today's rapidly evolving AI landscape, its previous features seem somewhat outdated. Last week, OpenAI unveiled GPT-4o, a more advanced AI with bidirectional conversational capabilities that surpass Alexa, such as real-time translation of conversations into different languages. Additionally, Google (GOOGL.US) introduced similar generative AI voice features for Gemini.

David Limp, Amazon's former Senior Vice President of Devices and Services, told The Verge in September last year that the company is considering charging for a more powerful version of Alexa.

Some view last week's announcement as a threat to Alexa and Siri. New York University professor Scott Galloway referred to these updates as "Alexa and Siri killers" in a recent podcast. Currently, many people use Alexa and Siri for basic tasks like setting timers or alarms and checking the weather.

Insiders revealed that the development of new AI chatbots in recent months has increased internal pressure on the department, which was once considered a favorite of Amazon's founder Jeff Bezos. However, since his departure, the department has been under strict profit constraints.

Three former employees pointed out that Bezos was initially very passionate about Alexa, considering it his pet project. Bezos' attention brought in more funding and reduced the immediate pressure for returns from these funds.

After Andy Jassy took over as CEO in 2021, following Bezos, changes occurred. Sources say that during the pandemic, Jassy was responsible for rationalizing Amazon's business, leading to Alexa losing some of its importance internally. According to an insider, Jassy has privately expressed dissatisfaction with the current capabilities of Alexa. For example, a source present at the time mentioned that Jassy, a sports enthusiast, once asked the voice assistant for real-time scores of a recent game, but Alexa was unable to provide an answer easily found online, leading Jassy to openly express frustration Amazon pointed out in an interview that the company released its annual shareholder letter last month. Jeff Bezos mentioned in the letter that the company is building "a large number of GenAI applications covering all of Amazon's consumer businesses," and added that this includes "smarter and more capable Alexa."

The team's current task is to develop Alexa into a relevant device, establish a foothold in the new artificial intelligence competition, and prove that Amazon's investment of resources and manpower is reasonable. According to three sources, the team has undergone a large-scale restructuring, with most personnel transferred to the Artificial General Intelligence (AGI) team. Other sources indicate that the team of thousands of employees working on Alexa has become bloated.

Amazon stated that as of 2023, the company has sold over 500 million Alexa-enabled devices, firmly establishing itself among consumers.

Is Alexa's Arrival Premature?

Apple, Amazon, and Google have all been early adopters of artificial intelligence voice assistants, but the current cutting-edge GenAI wave can still further empower interactions to be more creative and closer to human behavior. According to foreign media reports, Apple Inc. (AAPL.US) is expected to introduce a more conversational Siri at its annual developer conference in June this year.

Former employees of the Alexa team believe it is a great idea, but it may be premature and difficult to change course.

Furthermore, recruiting artificial intelligence engineering talent is a challenge as OpenAI, Microsoft, and Google all draw from the same academic and technical talent pool. Additionally, the cost of executing generative artificial intelligence tasks is high due to the need for hardware and computing power. One source estimates the cost of using Alexa at 2 cents per query and internally proposed a price of $20. Another source believes it should be in the single-digit dollars range. OpenAI's ChatGPT advanced model charges $20 per month.

Despite this, they still point out the user potential of Alexa, as devices with Alexa installed are present in millions of households, presenting a huge opportunity. However, Alexa staff also acknowledge the risks, as Alexa has entered people's living rooms and kitchens, posing greater risks if it fails to understand commands or provides unreliable information, leading to more serious consequences.

Amazon has been battling the perception that the company lags behind in artificial intelligence. While it offers various artificial intelligence models on Amazon AWS, it currently does not have a leading large language model to compete with OpenAI, Google, or Meta (META.US). Amazon invested $2.75 billion to support the artificial intelligence startup Anthropic, its largest venture capital investment in its thirty-year history. Google also invested in Anthropic and established a partnership According to a knowledgeable source, Amazon will use its own large language model Titan in the Alexa upgrade.

Two sources familiar with Bezos revealed that Bezos is also one of the people concerned about Amazon's lag in the field of artificial intelligence. According to foreign media reports last week, Bezos is still "very concerned" about Amazon's strength in artificial intelligence. He has been emailing Amazon's executives, hoping to find out why some AI startups choose other cloud providers instead of AWS