Most people engage in self-talk rather than not. A study found that 96% of adults have inner conversations, and though speaking to oneself aloud is less frequent, 25% of adults admit to doing so. Many individuals talk to themselves as a part of their daily routines. So if you ever thought that you were alone in this, that is clearly not the case. Speaking out aloud, irrespective of the topic of conversation, is actually quite entertaining. But what if all of a sudden you hear someone responding? Not your inner voice, but a second party who is hovering around in your house. And what if the speaker sounds oddly human despite the fact that you are 100% sure that you live by yourself? Fear not, you are not having auditory hallucinations, neither is your residence haunted, but it is just Alexa stepping her game up. Amazon’s Alexa is about to be your newest and trusted conversational partner with its latest developments.
Amazon ft. generative AI
A majority of inventions and advancements that come into play in the contemporary world have generative AI as the common denominator. Tech companies far and wide have found out that collaborating with the hottest artist on the block is the easiest way to come on top of the game. Therefore, the incorporation of AI in any and all its forms in almost all projects in the realm of technology has become a canon event. This time around, Amazon is set to bring advanced generative AI capabilities to Alexa, its widely used voice assistant. They made this announcement during a press event at their new campus in Arlington, Virginia on 20 September 2023. These upgrades will not only apply to the latest Alexa devices but also to all Echo products dating back to 2014.
Titled as ‘Let’s Chat’, this feature aims to make Alexa sound more natural and responsive. Currently, when interacting with smart home devices, there can be a noticeable delay between asking a question and receiving an answer. ‘Let’s Chat’ aims to minimise this delay, making Alexa function more like a real person who responds promptly. It will also enable users to modify their questions mid-sentence, and Alexa will still understand and provide relevant responses.
The enhanced Alexa will be capable of resuming conversations without needing a wake word, responding faster, understanding user preferences, handling follow-up questions, and adjusting its tone based on the topic. Not only will it have the ability to be conversational, but it will also be opinionated. If you think that your favourite singer deserved a Grammy and was done dirty by the award show, Alexa will either agree with you or dispute your claim, whilst providing substantial reasons. This will enable users to have fully fledged meaningful conversations that will not only influence the way in which humans think but will also provide Alexa with a portal to microanalyse the way in which your mind functions and scrutinise your thought patterns. Making it smarter and relatable, simultaneously.
For instance, as mentioned in CNN, during a live demonstration, Dave Limp, the Senior VP of Devices and Services at Amazon, inquired about his preferred college football team without explicitly mentioning the team’s name. He mentioned that he had previously informed Alexa about his favourite team, and Alexa remembered this information. Depending on whether his favourite team won or lost, Alexa reacted with either a cheerful or empathetic response. It is important to note that generative AI, in this context, refers to AI that can generate new content like text and images in response to user prompts in real time, making interactions with Alexa feel more like conversing with an actual human being.
Amazon emphasises the need for Alexa to be aware of real-time events and information because it operates in the physical world rather than just on the internet. For instance, when you inquire about the timing of the next BTS concert, Alexa must be capable of accessing the internet for current details and provide you with a fact sheet about their plans for military enlistment, along with a detailed analysis of the solo activities of each member.
Additionally, Amazon is in the process of developing a “speech-to-speech” model for Alexa. This advancement aims to enable Alexa to display human-like characteristics, including expressions like laughter and responses/filler words such as “uh-huh” in the course of conversations.
Amazon’s decision to introduce these updates is more of a response to external conditions rather than internal ones. To be considered worthy amongst the vying tech companies, it is vital to keep up with the rapid advancement of conversational AI tools, which has intensified the competition in the tech industry and fundamentally changed consumers’ expectations for their technology products. However, Amazon has not disclosed information on when these updates will be integrated into their products.
The future of voice assistants
Ensuing Amazon’s progress, ‘if Alexa could do it, so can we’ will be an attitude that will be adopted rapidly by many companies that use voice assistants. For instance, Qualcomm (an MNC known for semiconductors and wireless telecommunication products) envisions a future where cars offer voice assistants as advanced as ChatGPT. These voice assistants would have natural conversations with drivers and passengers, connected to the cloud for seamless interactions.
While we’re already familiar with basic voice commands like “navigate home” or “play the radio,” Qualcomm anticipates AI systems in cars that can handle more complex requests. While this could still be a far cry away, an example would be how you could ask your car to take you on a scenic route and point out beautiful spots, or even order food from your favourite restaurant to arrive just as you pull into your driveway. Although it may take some time to fully trust voice assistants with these tasks, it reflects the direction tech companies believe our interactions with AI in cars will go in the near future. And this is just one of many directions in which these innovations could go.
(Sandunlekha Ekanayake)