Why Amazon's New AI-Powered Echo Smart Assistants Raise Privacy Concerns

Amazon's Alexa devices are getting easier to talk to, but the new technology brings privacy concerns. 

The company has revealed a significant AI enhancement for Alexa, impacting all Echo products since 2014. This software upgrade allows Alexa to carry on conversations without a wake word. It boosts Alexa's ability to swiftly respond, understand user preferences, and manage follow-up questions. Moreover, Alexa will now be able to change its tone based on the topic of conversation and even share opinions, like talking about movies that should have won an Oscar. However, the use of AI is also raising red flags. 

"Conversational AI could be used by bad actors in a variety of contexts," Alyssa Devine, founder of Purple Fox Legal and a lawyer who works in privacy and AI, told Lifewire in an email interview. "The data collected by conversational AI may be used to create detailed profiles of users that can be sold for advertising purposes or monetized in another way without users' explicit consent. Even worse, this data could potentially be used by bad actors to commit fraud or identity theft."

A Chattier Alexa

The updated Alexa boasts a voice that mirrors human speech and can engage in fluid conversations without needing a wake word. With every interaction, it grows more familiar with its users.

Much like ChatGPT and other generative AI platforms, Alexa can now craft messages for users and dispatch them as requested.

"We've also focused on reducing latency so conversations flow naturally, without pause, and responses are the right length for voice—not the equivalent of listening to paragraph after paragraph read aloud," Daniel Rausch, the company's vice president for Alexa and Fire TV wrote on the company's website. "When you ask for the latest on a trending news story, you get a succinct response with only the most relevant information. If you want to know more, you can follow up." 

But Alexa's new abilities may open a window for data leaks. It will be important to read the fine print and understand how Amazon is using your interactions with Alexa and Echo devices to train and build its models and conversational speech recognition engine, Melissa Bischoping, the director of endpoint security research at the cybersecurity firm Tanium, said in an email to Lifewire.

"For some users, this may present privacy concerns that their conversations would be heard or reviewed by Amazon researchers or—in a worst-case scenario—potentially disclosed via a conversational form of prompt injection in the future," she added. "Amazon hasn't provided details on the guardrails in place to protect privacy, so it will be interesting to see what details and policies emerge on this topic."

Evan Macmillan, the CEO of Gridspace, a firm that provides AI voice agents for companies, downplayed privacy concerns. He pointed out that AI agents can be instructed to forget personal data. But, he said, as conversational AI interfaces grow in popularity and become the digital front door to more services, more bad guys will try to break in. 

"Companies should continue to be vigilant about authentication and intentional about what conversational AI interfaces have access to," he added.

How to Protect Yourself

Most AI-enhanced devices like Amazon's Echo use a mobile app to verify subscription and user information and update software, Monique Becenti, director of product marketing at the cybersecurity Zimperium, said in an email to Lifewire. She suggested that before downloading apps, check the app store to verify how the app will use your data and if your data will be linked to your identity. 

"Additionally, enable the minimum level for permissions in the app settings to limit the time the app can access personal devices," she added. "Like any software, conversational AI may have vulnerabilities that attackers can exploit."

If you're really concerned about the privacy of an AI assistant in your home, the best protection would be to not use one, Bischoping noted. She said that you should assume your conversations are being recorded and analyzed. 

"As with anything in the world, if you're not paying for the product, you are the product," she added. "And you should be mindful of how and where you decide to implement this kind of technology."