From voice synthesis to fertility monitoring, listed here are some truly useful AI merchandise at CES

“AI” was all over the place this year at CES; you couldn’t swing a badge with out hitting some firm claiming generative AI was going to revolutionize your sleep, enamel, or enterprise. However just a few functions of machine studying stood out as genuinely useful or shocking — listed here are just a few examples of AI that may truly do some good.

The entire concept that AI may not be a complete crimson flag occurred to me once I chatted with Whispp at a press occasion. This small staff is engaged on voicing the unvoiced, which means individuals who have bother talking usually as a consequence of a situation or sickness.

The title refers to situations the place an individual is ready to kind phrases, however whose vocal cords have been taken out of the image by, for instance, throat most cancers or an damage. These of us can whisper simply effective, however not converse — usually having to depend on a decidedly last-century digital voice field. So Whispp’s first massive characteristic is to synthesize their voices and switch these whispers into full speech.

The artificial voice is created by related means as different platforms — just a few previous recordings of somebody and you may customise a voice mannequin to sound sufficiently like them. Whispp’s fundamental problem, it appears, was to construct a speech recognition mannequin that labored properly with whispers and different affected speech. Curiously, this additionally works for individuals who stutter, since for no matter cause whispering usually reduces stuttering by an enormous quantity. You can read more about Whispp in our post about it here.

Across the nook from Whispp I bumped into the extraordinarily cheerful girls of Louise, a French startup targeted on fertility monitoring and recommendation for each women and men seeking to enhance their possibilities of conceiving.

Louise is essentially a B2B affair, working with hospitals and fertility clinics to investigate affected person information. It makes use of machine studying as a sign detector, sorting by means of hundreds of knowledge factors from exams and surveys and searching for biomarkers that would give perception into the complicated means of enhancing fertility. AI is nice at discovering refined correlations or patterns in massive collections of knowledge, and fertility is unquestionably an space that would profit from extra of that.

Picture Credit: Louise

The corporate was truly at CES to advertise its new app, Olly, which is its first B2C providing: a start-to-finish “fertility journey” app for each ladies and men (whose half within the course of is usually missed), from resolution to go for it all through success. It tracks appointments, presents documentation on drugs and methods, and so forth. And the icon is a cute little chick. Olly is deliberate for international launch on February 14.

The rabbit r1 acquired a good quantity of hype at CES, as a candy-colored pocket AI assistant ought to. However whereas it’s anybody’s guess whether or not the corporate will survive lengthy sufficient to breed (the r2, one presumes), this little doodad’s capabilities may truly be extra useful to folks with imaginative and prescient impairments than sighted of us who simply don’t wish to pull out their telephones.

Alexa, Siri, and different voice assistants have been transformative for numerous folks for whom navigating a smartphone or desktop laptop is a ache because of the basically graphical person interface. With the ability to simply converse and get fundamental information just like the climate, the information, and so forth is a large enabler.

The issue is that these so-called assistants couldn’t do a lot outdoors some strictly outlined duties and APIs. So that you may have the ability to discover out when a flight is leaving, however you possibly can’t rebook it. You’ll be able to e book a automotive however not customise the trip for accessibility. Search for trip locations however not hire a cabin on the seaside there. Stuff like that. The r1 is constructed to have the ability to not simply do fundamental assistant queries by means of a ChatGPT-like voice interface, but additionally function any regular cellphone or internet app.


The demo mannequin on Rabbit’s web site.

If the machine and repair match the corporate’s claims, the r1 could possibly be a helpful helper certainly for anybody who has bother interacting with a conventional laptop. In case you can speak, you will get issues executed — and if you happen to use Whispp, you don’t even want to speak!

Elder care is one other space the place among the frequent criticisms of generative AI fall quick. I don’t assume anybody ought to should depend on a pc for companionship, however that doesn’t imply that the computer systems we already should work together with couldn’t be slightly nicer about it. Though I’ve made it clear (at nice size) that I don’t assume they need to faux to be folks, they will nonetheless be pleasant and useful.

ElliQ makes gadgets (“robots”) for locations like assistive care services, the place having a gadget within the room that may remind sufferers of issues or ask them how they’re is a win-win addition. The most recent machine makes use of a big language mannequin to provide extra pure dialog, and like ChatGPT and others, it could actually discuss absolutely anything. Many older of us are starved for dialog or battle to know new developments in tech or the information, so this could possibly be a genuinely great way for them to remain engaged. Plus it provides a component of care and security, in a position to hear the particular person in the event that they name for assist, or relay requests to caregivers, or simply examine in.

That’s not all the great functions of AI at CES, however it’s sufficient of a pattern to point out you that whereas it could be the following massive hyped factor, that doesn’t imply there aren’t any good methods to use machine studying in on a regular basis life.

Read more about CES 2024 on TechCrunch