Chatbot-Like Siri Patent Includes Intelligent Image, Video, and Audio Recognition Within Messages

A patent application published by the United States Patent and Trademark Office today details a new Apple service where users could make inquiries and talk with the company's AI assistant Siri through Messages (via AppleInsider). The new patent is similar to a filing the USPTO published late last year, but now includes deeper integration with audio, video, and image files.

Similar to chatbots in Facebook Messenger and other texting services, Apple's patent describes a Siri that could perform her current duties without the user having to speak aloud, which could be helpful in certain public situations.

The "Intelligent Automated Assistant in a Messaging Environment" could respond to text, audio, images, and video when sent to it by the user, which Apple said would result in "a richer interactive experience between a user and a digital assistant." The patent gives a few examples of a conversation held between Siri and a user in Messages, with the user asking questions regarding calorie content in food, upcoming meetings, and even asking Siri to text a friend.


Interesting applications include a thread where a user texts Siri a picture of a car or a bottle of wine, and Siri sees the images and can intelligently respond to the user's inquiries about them. For the car, the user asks Siri for details on pricing for a specific model using only an image, and Siri searches the internet and returns the relevant MSRP information.

The bottle of wine image is used as an example to show Siri's memory functions, where a user asks Siri to remember their favorite wine, which she can resurface at a later date. Siri sees the wine image, reads the label, and can then respond to a user's question in text format about the brand and even year it was made.


Other image-related inquiries include "Where is this place?" and "What insect is this?", to which Siri would respond "This is the country Algeria" and "This is an earwig," respectively. Audio and video could also be recognized by Siri, including simple Shazam-like questions related to songs and the content of shared videos.

Apple points out in its patent that thanks to the chronological format of texting, users would be able to "review previous interactions" with Siri, unlike how current Siri conversations disappear immediately after they conclude. Subsequently, Siri would be able to use that history to become smarter and "define a wider range of tasks."
The messaging platform can enable multiple modes of input (e.g., text, audio, images, video, etc.) to be sent and received. As described herein, this can increase the functionality and capabilities of the digital assistant, thereby providing a richer interactive experience between a user and a digital assistant.

A digital assistant in a message environment can thus enable greater accessibility to the digital assistant. In particular, the digital assistant can be accessible in noisy environments or in environments where audio output is not desired (e.g., the library). Moreover, the chronological format enables a user to conveniently review previous interactions with the digital assistant and utilize the contextual history associated with the previous interactions to define a wider range of tasks.
The patent includes a description where Siri would be "a participant in a multi-party conversation," allowing group chats to use Apple's AI simultaneously. Apple gives an example where one user asks Siri to list nearby Chinese restaurants to begin making the group's dinner plans, and then another user responds by asking Siri to whittle down the list to only include the cheapest places. One user's personal Siri can even be asked to remind other participants of the upcoming dinner.


Apple is believed to be working on an "enhanced Siri" that might launch in iOS 11 this fall, but the exact specifications as to what would make the new Siri "enhanced" have never been divulged. A questionable rumor in March stated that deep Siri integration is coming to Messages in iOS 11, but the source of the news -- The Verifier -- doesn't have a previous track record of reporting accurate rumors.

Chatbots are certainly growing in popularity so it wouldn't be too surprising if Apple introduced some kind of text-based Siri interface, particularly considering the multiple patents the company has published on the topic. Still, as with all patents it's best to look at Apple's new filing as an intriguing insight into what the company might be working on for the future, rather than proof of an impending launch.



Top Rated Comments

(View all)
Avatar
30 months ago

So instead of fixing Siri's speech interface Apple wants you to type in queries...

...so different than Googling.


bc lots of people prefer to not talking in to a phone in public places or quite places.
Rating: 6 Votes
Avatar
30 months ago

So instead of fixing Siri's speech interface Apple wants you to type in queries...

...so different than Googling.

.....are you asserting that somehow working on one feature means that the core of Siri isn't being worked on?
Rating: 5 Votes
Avatar
30 months ago
. . . . . . . . [image of car]

What should I do with it?

. . . . . . . . How much does it cost?

OK, now playing songs by Lady Gaga.
Rating: 3 Votes
Avatar
30 months ago
So instead of fixing Siri's speech interface Apple wants you to type in queries...

...so different than Googling.
Rating: 3 Votes
Avatar
30 months ago
Low-hanging Siri jokes aside, this seems interesting. I welcome any traction on Siri development. And while today's Siri makes the thought of Siri inquiries via photo or video seems laughable, the next update to Siri could make this concept more plausible.

As one who tries to interact with Siri as much as possible, this Messages concept could be nice for a couple of reasons.

1. An archived Siri thread.
I'm often frustrated when after conducting a Siri inquiry and then leaving the Siri window to (say) check a link, I can't get back to my initial Siri interaction to review, continue or amend. Even a standalone Siri "app" could be interesting. Could offer a simple/intuitive way for users to "get back" to their Siri-ing via Home button double-click to reveal a Siri slide in the app switcher.

2. The ability to access Siri via text input.
There are definitely scenarios where accessing Siri via voice isn't always desirable. Being able to interact with her via text is a welcome option.

Will be interesting to see what Apple has up their sleeve regarding Siri. Will
Be nice if they update us on Siri dev at WWDC. While it can never been good enough, any development/advancement of Siri will be welcomed.
Rating: 2 Votes
Avatar
30 months ago

A cursory glance through what Google regurgitated showed three schematics that just say 'carrier,' so this does seem to be outside the norm. What it indicates? Who knows, probably nothing. :p

Apple buys T-Mobile confirmed obviously!!
Rating: 1 Votes
Avatar
30 months ago

I've never noticed this before, but do they always put a real carrier name on these patent mockups? I thought they usually leave them intentionally vague. Looks like Apple is a fan of T-Mobile as well!


Well, Apple's definitely not a fan of AT&T and T-Mobile has a reputation of being the cool carrier that tries to be more like the best European/Asian wireless carriers, giving customers more choices, features that you would expect in 2017 like seamless/inexpensive international data roaming, and a better value.
Rating: 1 Votes
Avatar
30 months ago

So instead of fixing Siri's speech interface Apple wants you to type in queries...

...so different than Googling.

I guess the main difference would be Siri acting on local information like calendars, contacts, and Apps. I think this is mainly useful for when you want to fetch that kind of relevant info and not have to ask for it again if you need to exit the screen and do something else.
Rating: 1 Votes
Avatar
30 months ago

And most people prefer google search to Siri search so...


So now we have 2 choices (and many others) instead of google, choices are good.
Rating: 1 Votes
[ Read All Comments ]