7 New Developments in AI

by Jonathan Tarud
Blog Post

Artificial Intelligence (AI) is one of the major developments of our time. Machine learning and the implications that go with it are shaking up many aspects of how we do things, allowing us to deploy AI where we previously used a human, or a more inefficient process.

Sometimes this is to the consternation of people, particularly those who worry about AI taking human jobs, or perhaps the sci-fi scenario of AI being intelligent and organized enough to overrule humans.

One thing we do know is that we’ve probably only scratched the surface in terms of what is possible. As Oracle EVP and head of applications, Steve Miranda, said at a recent event; “Two years from now, we’ll probably be talking about a whole new set of things in this category that probably none of us is even thinking about today.”

Well, that’s an exciting thought, considering how far we’ve already come! Here are some new developments in AI, which demonstrate how the technology is advancing:

#1. Robot learns by observing

The mechanism through which AI “learns” is generally through training by humans or “machine learning,” where the bot learns by processing data. For example, a bot might observe that you seem to go to the same place at the same time every day, and it may start to automatically look for traffic and weather conditions, to provide you with an estimated driving time.

A recent groundbreaking development in AI has been robots learning through observing the actions of humans. Nvidia demonstrated a robot that performs tasks in a real-world setting by watching how they are done, a different and more hands-off mechanism from how robots are usually trained.

If robots can learn through observing demonstrations, this has implications, particularly for the workplace and taking care of physical tasks. Perhaps robots of the future will be in homes, observing how household tasks are performed and taking care of those?

In another development along similar lines, a bot program called AlphaGo taught itself advanced strategies for playing the game Go, with no training from humans. This is further highlighting a growing trend of AI that is able to be independent of human knowledge.

Developments in AI

#2. Robot caregivers to fill shortfall

How would you feel about being cared for by a robot nurse, or your elderly relatives being cared for by robot caregivers? Many countries throughout the world are heading towards a crisis in terms of having enough carers for aging populations. Particularly as the large baby boomer generation reach their elderly years, the shortage is predicted to be more pronounced.

Artificial intelligence is being developed to step in and make up for the shortfall. The Japanese Government in particular are working on increasing acceptance of technology filling in for human nursing and caregiving roles.

Japan is facing a predicted shortfall of 370,000 caregivers by 2025, and developers are focusing their attention on simple applications of AI. For example, a robot might help a person to get out of bed, or it might predict when a patient is going to need to use the toilet.

Potential resistance to help from a robot is one of the issues researchers are working on. The next research priorities include wearable mobility aid devices and technology that guides people to the toilet at what it predicts is the right time.

#3. AI beer brewers

What if the perfect beer could be brewed using the help of AI? Okay, “perfect” is going to very much be in the eye of the beholder, but AI is being used by IntelligentX to take customer feedback into account as much as possible. So technically, the beer is a product of customer experience, AI and skilled brewers.

Basically, it works by using an algorithm that is behind a Facebook Messenger bot. The bot takes the customer feedback and passes it on to the humans who are actually brewing the beer. The technology facilitates brewers receiving that feedback quicker than they ever did before.

The company places codes on bottles which direct customers to the bot. They are then asked a series of questions, the answers to which are interpreted by the algorithm. Feedback is accumulated to spot trends and inform the brewing process.

#4. AI-based cybersecurity

Cybersecurity has been a hot issue ever since it became necessary. As technology evolves, so do potential threats to information and networks.

There has been increased demand for AI solutions to boost cybersecurity. Professionals are hoping it will accelerate incident detection, improve incident response, identify and communicate risk, and generally help them to maintain optimum situational awareness.

Palo Alto Networks recently introduced Magnifier, a behavioral analytics AI solution. It models network behavior by using structured and unstructured machine learning, to improve threat detection.

Also this year, Google’s parent company Alphabet introduced Chronicle, a cybersecurity intelligence platform. Chronicle is a powerhouse for cybersecurity data, allowing for rapid search and discovery. The idea is that security teams already have the information they need within their systems, but it is often hidden among the millions of data points. Machine learning advanced search capabilities are the driver for more rapid search.

#5. AI diagnostics for x-rays

Medical technology is a field ripe for innovation from AI. Areas such as diagnostics traditionally rely on a human operator being able to read and interpret tests or imaging results. This naturally creates some kind of lag in processing, and leaves open the possibility for human error.

There are major challenges in the area of AI for diagnostics. For example, the AI must be taught to correctly interpret results under human supervision, and it is difficult to teach the identification of rare pathologies, due to a shortage of images.

A recent development has essentially “used machine learning, to do machine learning,” by using computer-generated x-rays to augment AI training.

“We are creating simulated X-rays that reflect certain rare conditions so that we can combine them with real X-rays to have a sufficiently large database to train the neural networks to identify these conditions in other X-rays.”Shahrokh Valaee

This development brings the idea of AI actually taking the diagnostics role even closer.

Developments in AI

#6. AI in smartphone apps

AI is making an appearance in a broad range of smartphone apps that are designed for regular consumers. Gartner predicts that, by 2022, 80% of smartphones will be equipped with on-device AI capabilities (compared with 10% right now).

This makes AI a key opportunity for developers of all types of apps. Here are just a few that are currently in use:

  • Google Assistant – You can access your assistant by holding down the home button on your Android phone, or saying aloud “okay Google.” From there you can send messages, check appointments, play music, and a host of other things hands-free.
  • Socratic – Math help is here! Socratic is a smart tutoring app which can explain how to solve problems through analyzing a picture of the math problem.
  • Microsoft Pix – Everyone wants to be able to take and share the perfect photo. Microsoft Pix helps by capturing ten frames per shutter click, using AI to select the best three, then deleting the rest, saving you storage space.

#7. AI in FinTech

FinTech has been an area to see a lot of disruptive technology in the last decade. Traditional financial institutions have found themselves much less popular, as other solutions, such as apps have popped up. AI is another disruptor in the sector.

AI is able to reduce processing times – for example, your bank probably has an app which allows you to photograph a check for deposit. The funds are often available immediately, in part due to AI being able to read the check. This eliminates the need for a human operator to accurately read and deposit the check.

Fraud detection is another which AI is helping with. For example, Pixmettle is developing enterprise-level AI tools to help flag things like duplicate expenses and corporate policy violations.

Chatbots are also now widely in use. Many banking apps use them as part of their customer service suite, while there are apps that have specifically been developed to connect financial accounts with Facebook Messenger (such as Trim), allowing users to ask questions via the app, make cancellations, or get reports.

Of course, FinTech is also ripe for cybersecurity AI, as mentioned earlier. AI is scalable, and able to rapidly analyze large amounts of data.

Final thoughts

Artificial Intelligence represents a huge opportunity across virtually every sector. It has already proven to be disruptive, but it is anticipated that it will be much more widespread over the next few years.

How can your industry use AI to make improvements? How will you ensure that your company is competitive and makes the most of new technologies? AI is an area to watch, and to think about the applications for your own niche.

Koombea builds state-of-the-art apps for companies. Talk to us today about how we can help you.

by Jonathan Tarud
Blog Post