18 Cool, Creepy, and Cutting-edge Applications of Artificial Intelligence (featuring UC startups)

Everyone has heard the discussions of artificial intelligence, whether it is dangerous, how it’s going to take over the world, and how it will automate a large number of jobs in the not-so-distant future. Believe it or not, we’re already surrounded by artificial intelligence, though we may not recognize it. There are over 50 artificial intelligence startups at UC, so we thought it was time to delve into the world of AI and look at the cool, creepy, and cutting-edge applications of today.

Email

Believe it or not, but we have much to thank AI when it comes to our inboxes. AI is the reason our spam filters work: an algorithm is trained and learns to recognize which emails are spam and which emails are not spam. While it’s incredibly satisfying mass deleting in one fell swoop all the spam that comes to your inbox, we don’t stop to think that the very act is enabled by AI.

Gmail users might be divided on the next use of AI: Google divides Gmail inboxes into Primary, Social, and Promotions – again, by using artificial intelligence. Some people love it, some people hate it.

Entertainment

Name the big online television providers, and chances are 100% that they are using AI to help you discover and watch even more series, seasons, and episodes. Using a massive pool of data from their users, giants such as Netflix, Hulu, and Amazon Video are using algorithms to predict other shows and seasons in which you might be interested. Going a step further, they give you recommendations based on your viewing history, browsing history, or purchase history.

Retail

In a similar fashion, online retailers make recommendations based on your viewing history, browsing history, purchase history, and more. Amazon, Walmart, eBay, and most of the major e-retailers have those well-known sections on their websites: “customers who bought this also bought…” or “inspired by your browsing” or “customers who looked at this item also looked at…” Everything is based on data and designed to increase purchases through helpful discovery.

We’re at a point where as buyers, we expect online retailers to make recommendations, suggestions, and help us discover relevant products. What’s more cutting edge from a retail point of view is Heineken’s use of AI. The beer company uses data collection and analysis for its supply chain as well as point-of-sale data through a partnership with Walmart to collect data on customer purchases. None of that is unusual – but what they did with that data as well as data from social media is incredible. Knowing that a lot of consumers drank Heineken during the EUFA Champions League Final, the company created the Heineken Banternator, the world’ first AI football commentator that responded to Heineken’s Twitter followers with “relevant football banter during every aspect of the entire game.”

Healthcare and Medicine

AI in healthcare and medicine has huge potential as well as potential to positively impact lives. Catalia Health helps patients with chronic illnesses to manage their illnesses. Catalia Health created Mabu, a wellness coach that’s an AI who provides tailored conversations to each patient that evolve over time. C. Light Technologies is a UC startup that aids in the prognosis and monitoring of Multiple Sclerosis through non-invasive retina imaging. The technology uses algorithms to help diagnose and monitor MS based on novel eye motion data. Arterys is a clinical analytics platform for medical imaging: consider it the AI assistant for radiologists. It makes radiologists’ job faster and easier, and outcomes more accurate. RADLogics is doing similar work and provides its Virtual Resident, a platform that supports the analysis of CT, x-ray, MRI, and ultrasound scans and uses algorithms to incorporate search, measurement, and other findings into a preliminary report.

Singlera Genomics creates non-invasive genetic tests to diagnose diseases, including cancer, and creates personalized cancer treatments. They help with the early detection of cancer: data suggest detection four years before the earliest conventional diagnosis. In addition to their proprietary medical techniques, they use AI to create an empirical classifier, in short, a predictor of whether or not someone has cancer. Arguably the equivalent for the heart market, Eko Devices is working on using algorithms to analyze heart sound and EKG data from its hardware devices such as stethoscopes. It wants to help prevent heart attacks and detect murmurs or anomalies.

Beauty

On a mission to make every woman feel beautiful in their skin, AllOfUs Cosmetics helps buyers find the exact foundation for their skin tone. Sit down in one of their photobooths, take a photo, then their software uses AI to conduct a color analysis of your skin, producing a recommendation for which foundation is best for you.

Finance

Just about every credit card company uses data analysis to predict who should or should not be offered a credit card and how much credit should be offered. Multiple variables are considered such as income, age, job status, pervious credit history, etc. But credit card companies are getting smart and have come up with uses of AI to influence customers and their experience, budgeting, and savings. Hailed for and evaluated by their rewards programs, credit card companies are starting to personalize reward offers. There’s a shift from the typical categories: “the traveler” or “the businessman” to customized reward experiences based one each customer. Think a step further: it’s not hard for a credit card company to know your travel preferences if you purchase all of your travel on your credit card. Alas an AI personal travel concierge called Mezi was born, and recently acquired by American Express. Along similar lines, the Bank of America launched Erica, an AI personal assistant that shows you what you spend where, how much, and suggests tips on how to help you save money.

The darker side of AI in finance relates to fraud detection. Companies are using AI to predict when to classify a transaction as fraudulent for a given customer. Smart banks recognize when a customer has bought a plane ticket to say, Arizona, then starts making transactions one day from Phoenix. Are the Phoenix transactions fraudulent? Not if those are the dates for which you booked your trip! Well programmed AIs figure that out. Of course, not all banks have well programmed AIs and the result is that you find yourself in Phoenix with a credit card that has been automatically cancelled and you’re likely quite unhappy with your bank.

Children’s Toys

Barbie has ventured into the world of AI and natural language processing and the result, I find, is creepy. They launched Hello Barbie, a doll that responds to what a child says. The doll has a microphone concealed in her necklace that records what anyone says to her and transmits it to ToyTalk’s servers. The input is analyzed, a “correct” response selected, and a command sent to Barbie to deliver said response. Hello Barbie stores what she hears and what she answers for use later, so if you tell her your favorite color is black, Barbie will know!

Illegal Fishing

I have to give Google credit for what they’re doing with AI to help prevent illegal fishing. Google allocated some of their engineers to put together data from their maps, the Automatic Identification System for shipping, and real-time ship location data to create algorithms that identify illegal fishing activity. The algorithms can predict a ship’s course, identify its reason for being at sea, produce heat maps of fishing areas, and identify suspicious activities, including vessels in a UNESCO heritage marine site.

Legal Services

Any entrepreneur or executive worried about protecting their IP can breathe a little easier thanks to DataNovo. The company uses algorithms to identify patents that have been potentially infringed upon, prior patents that are the most relevant to a given patent, and patents that may be infringed upon given a six-word product description. So you can either more quickly and easily protect your IP or prevent yourself from infringing on a patent, especially if you weren’t aware of it. Consider it AI applied to the entire database of US patents.

Education

The rise of the edtech industry has produced a simultaneous amount of data around education and learning. At the K12 level, schools are using data analytics and algorithms to predict whether a students is on a path of becoming at-risk or worse, if they’re on a path to dropping out. This empowers teachers and schools to take actions to help prevent dropouts and ultimately help the future of their students. While there’s no clear leader in the space, there are schools and districts leading the way by using data from their various assessment and administration software.

Supply Chain

Anything that needs to be optimized and has a variety of variable inputs can be a good candidate for AI. DeepVu is applying AI to the supply chain to optimize the chain, maximize margins, and minimize risks. DeepVu doesn’t just use company data to create optimizations: it also uses context variables such as commodity prices, GDP, the weather, gas prices, and more. It even helps to predict supplier pricing including hedging on commodities.

Music

You might be surprised to learn that AI has been in the music industry for quite a while already. In case you might be wondering if any of the recent hits were made by an AI, the answer is no: no AI has yet created an industry-toppling hit. However, AI is enabling musicians, helping with the creative process, and helping achieve music visions.

One minor human frustration that AI has solved in the music world is the problem of hearing a song but you just can’t quite remember its name or who sings it. Thanks to apps such as Shazam (by UC alumnus Chris Barton) or Google’s Now Playing, we can find relief in knowing exactly what that song is and its artist.

Autonomous Vehicles

Finally, one of the hot topics in AI is the autonomous vehicles: will a vehicle become smart enough to drive on the road safely, even better than humans? Deepscale, a UC startup who just raised $15M, develops perception for autonomous vehicles. The startup is focusing on algorithm optimization and considers both speed and energy, not just improving accuracy of the vision technology. In short, if the vision is good but slow and consumes enormous amounts of energy, there’s no market fit.

From fishing to our inboxes to the music we listen to, AI is already all around us, and this is literally just the beginning. It’s no surprise that the technology is expected to grow and will touch every industry in our economy as well as the structure and make-up of our society. Naturally, using AI to create value and profitable businesses will occur, but our challenges lie in areas that enter the ethical realm such as how to program a self-driving car when it’s faced with probabilities of killing drivers or passengers or pedestrians, other drivers, or living beings not in the vehicle. Or to what extent do we use AI to predict crime when doing so encroaches on the freedom and rights of citizens? We’ll see where we end up; ultimately, it’s our responsibility to actively steer our society and technology where we do and don’t want it to be.