Concerned about privacy? Your Amazon Echo devices will get creepier next week, when all Alexa conversations will be sent to the cloud after a policy change seemingly intended to feed data to the company’s souped-up generative AI (genAI) revamp.

The policy change – which was announced in a recent email to customers – will see Echo users blocked from enabling the ‘Do Not Send Voice Recordings’ option, which preserved privacy by forcing the devices to process their commands on the device.

From 28 March, however, this option will be disabled and replaced with a new option that only requires Amazon to delete the recordings from the cloud – eventually.

This means recordings of whatever you say to Alexa will be stored on Amazon’s systems long enough for it to be used to train the generative AI (genAI) features Amazon recently announced as part of its upcoming reboot of Alexa, called Alexa Plus.

“As we continue to expand Alexa’s capabilities with generative AI features that rely on the processing power of Amazon’s secure cloud,” WIRED quotes Amazon’s email as saying of the Do Not Send option, “we have decided to no longer support this feature”.

Bringing genAI to the home

The policy change is part of the company’s effort to revitalise Alexa, which Amazon has reportedly managed to get into over 600 million devices worldwide but has struggled to monetise as genAI puts similar voice-based assistants into all kinds of products.

Amazon is invested in maximising the capabilities of Alexa+, which debuted weeks ago and with plans to be rolled out to existing devices and baked into a range of new Echo home devices set to be released later this year.

By tapping new features called ‘experts’, Amazon CEO Andy Jassy recently explained, Alexa+ “is going to be the first chatbot that is not only highly intelligent to answer questions, but she can do so many things for you.”

“It really is the first big, large scale, practical use of generative AI that consumers are going to be able to see and use naturally… you can do all the things that you used to do, but every single one of those functions is better.”

Amazon’s echo devices will face off against rivals from Google – which is integrating its Gemini genAI into its Nest devices – and new Apple devices such as its smart doorbell and the HomePad smart display, to be powered by the new homeOS operating system.

Your voice will be assimilated

As with all things genAI, adding and improving functionality leans heavily on training – which requires ever larger quantities of data be siphoned into cloud services.

Sourcing that data, it appears, has forced Amazon to back away from a simple fact: the whole reason Echo devices used to process recordings on the device was to allay consumers’ privacy concerns – which are now being discarded in the name of ‘better’.

In many ways, it’s natural for Amazon to try to increase use of cloud-based genAI services, since its multi billion-dollar Amazon Web Services (AWS) business and Bedrock genAI platform anchor offerings from hundreds of software companies.

Yet diverting customers’ audio recordings to that cloud put Amazon’s privacy policy in opposition to that of Apple – which has worked to keep processing of user commands and data on the device, playing to its own strengths as a device maker in the process.

Apple has not only fought to preserve encryption of data on its customers’ phones, but has built AI processing into chips such as its M4, which is powerful enough that many AI queries can be processed without sending data to the cloud.

It also released a new feature, Private Cloud Compute, that extends its device security to the cloud, for when the query can’t be completed without the support of a cloud AI server; that data, the company says, “is not stored or made accessible to Apple.”