Alexa+ Offers a Sneak Peek into the Next Generation of AI Assistants

Alexa+, built using a combination of Anthropic’s Claude (Amazon is an investor in Anthropic) and Amazon’s proprietary AI models, is now rolling out. With it, we may be witnessing the transformation of digital assistants into full-fledged digital companions.

Compared to its predecessor, Alexa+ boasts substantial improvements. Users can now converse without constantly repeating the wake phrase “Alexa.” It remembers past interactions, allowing for more natural, multi-turn dialogues— almost like conversing with a human assistant. Alexa+ can retain details such as frequent flyer numbers, dietary preferences, and other personal information, leveraging this knowledge for more personalized assistance.

Also, Alexa+ conversations and tasks can persist across multiple touchpoints and devices, such as Echo smart speakers, smartphones, and web browsers. This enables greater continuity of interactions; you can pick up from where you left off. Alexa+ can also summarize documents, emails, and notes and reference them in subsequent interactions.

On the home automation front, the upgraded version introduces more intuitive controls. For example, scene-based navigation on Fire TV is a welcome addition. Alexa+ also integrates more seamlessly with third-party apps, making it easier to shop, order an Uber, and more. Overall, this version is a significant leap forward from Alexa, which itself, until the beginning of the ChatGPT era, stood out as one of the more capable AI assistants.

Similar advancements can be expected from Siri and Google Assistant (which is now rebranded as Gemini). But while the original launch of voice assistants delivered many “wow” moments initially, consumer interest has largely plateaued over the past decade due to incremental improvements. With the integration of large language models (LLMs), the upgrade in voice assistant capabilities is far more than incremental.

That said, monetization continues to be a tricky challenge. Alexa+ is priced stand-alone at $19.99 per month but is also included in a Prime subscription, which currently costs just $14.99 per month. Yes, as of this writing, the Prime subscription, which also includes shipping, video, and music benefits (and now Alexa+) costs less than Alexa+ itself. This also raises a few interesting questions: Will Prime subscribers see a price hike? Or has Amazon determined that users won’t pay for Alexa+ separately and is instead using it to bolster the value of Prime? 

Another layer of complexity is the high cost of running LLMs; ChatGPT, for instance, reportedly loses money on each user query. If the LLM economics stay the same, scaling up usage won’t solve this issue. Instead, the companies will actually lose more money with greater adoption, though advancements of the kinds announced by AI firms like DeepSeek may eventually reduce these costs. Regardless of this cost calculus, Alexa+ represents a technological leap forward.

But as with earlier AI assistants, concerns about privacy, persistent tracking, and security remain or increase even more. The involvement of LLMs introduces new and unique risks such as AI hallucinations—errors that could lead to incorrect recommendations or unintended actions. And while these systems offer greater convenience, are we giving up control over our personal data in exchange?

Isaac Asimov famously proposed the Three Laws of Robotics to ensure human safety and control over robots. As we stand at the threshold of powerful AI companions, I similarly propose three guiding principles, Kompella’s AI Companion Code:

• Cause no harm. An AI assistant must prioritize user well-being and security, ensuring that its responses and actions do not mislead, violate privacy, or cause harm.
• Confess when you don’t know. An AI assistant must acknowledge uncertainty, admitting when it lacks knowledge rather than fabricating or misleading.
• Care for the vulnerable. An AI assistant must provide extra safeguards for vulnerable users, particularly children, ensuring interactions remain ethical, safe, and age-appropriate.

These principles (the 3 Cs) are increasingly important as AI assistants become more deeply integrated into our lives. The risk of hallucinations means we need AI responses to be transparent, while the growing reliance of children on AI assistants demands additional diligence and additional guardrails.

The futuristic AI assistants are here, but with their awesome power comes the need for appropriate accountability.