Technologies develop and evolve very quickly, and nothing illustrates this better than the rise of artificial intelligence (AI). Over the past few years, there’s been a lot of hype surrounding AI. In essence, AI represents a machine’s ability to ‘behave’ intelligently. With AI, complex tasks can be easily automated, which is why some are afraid this tech could soon make us obsolete. But the fear of automation isn’t stopping companies from pouring a lot of money into AI technology.
For instance, according to the Worldwide Semiannual Cognitive Artificial Intelligence Systems Spending Guide, created by the International Data Corporation (IDC), global spending on AI systems will reach $77.6 billion in 2022, compared to $24 billion in 2018. Banking and retail will be the biggest investors, followed by healthcare and manufacturing. The same report reveals that the US, Western Europe, and China will be the most significant regions when it comes to AI spending.
It seems that this technology has pushed itself into the spotlight, as companies across industries are using AI to improve their products and services. But is AI really that beneficial? A researcher specializing in deep learning, Jeremy Howard, claims that AI isn’t overhyped, and it certainly deserves all of its popularity. As Howard explains, AI and deep learning didn’t come out of anywhere. Instead, what we’re seeing today is the result of years and years of research, which required a lot of hard work. If you don’t think the same, just take a look at the recent innovations in this field.
AI-powered technology simplifies banking
Earlier this year, TD Bank announced the launch of its AI chatbot, called TD Clari. This solution was created to help TD Bank to reduce the number of call center requests and to make banking easier and more personalized. The chatbot uses the KAI Consumer Banking Platform developed by the fintech startup Kasisto. Thanks to Clari, which can be accessed through a mobile app, TD Bank’s customers will be provided with real-time insights into their transactions and other financial decisions.
For instance, if customers want to know how much they spent that month, they can simply ask Clari, and the chatbot will respond immediately. Besides transaction data, Clari will also inform its users about exchange rates and the nearest bank locations. Since it uses machine learning, this innovation will learn over time through interaction with the customer. Eventually, it could “better predict needs and prompt customers with deeper insights into their finances”, says Zor Gorelov, the CEO of Kasisto. Clari is currently available for iOS users, but the bank will soon be releasing an Android version, too.
AI translates thoughts into speech
But the finance industry isn’t the only one disrupted by AI. This technology could leave an indelible mark on the healthcare sector, too. By harnessing AI’s full potential, researchers from Columbia University have developed game-changing tech capable of translating brain signals directly into speech. Our voice is one of the things that connects us with our family and friends, which is why people who lose it due to an injury or disease can feel devastated. But Columbia’s neuro-engineers have found a way to restore people’s ability to speak and connect them with the surrounding world. The project, published in the journal Scientific Reports, explains how the team used a computer algorithm to understand human speech. The algorithm, named vocoder, is the same tech used in smart speakers such as Amazon’s Echo.
The researchers first measured the brain activity of epilepsy patients who were required to listen to four speakers that presented short stories. The team recorded the patients’ neural patterns and added the data to the vocoder, “which in turn produced synthesized speech”. As a result, the vocoder was able to produce a robotic voice that was understood by participants 75% of the time. The researchers hope that this discovery will enable them to develop an implant that could translate patients’ thoughts directly into speech.
Winning the war against counterfeiting with AI
AI could also help companies to identify fake products and prevent counterfeiting. This has become a serious issue, and it’s estimated that the global counterfeiting industry will be worth $1.82 trillion by 2020. This inspired the tech startup NeuroTags to develop an AI solution designed to identify fake products. NeuroTags consist of “algorithmically coupled tags”, with one visible on the product, while the other is kept inside the product. What makes this technology better than existing RFID tags is the fact that RFIDs are easy to replicate and are also quite expensive. NeuroTags are a better alternative.
While the open tag allows anyone with a smartphone to scan the item and get the product’s authenticity information, the protected tag, kept under the product’s seal, is available to customers only after they buy the product. By scanning this tag, customers automatically become registered as the product’s owner and get easy access to loyalty and referral points. Both tags are connected and protected by algorithms in an AI-based server. When the protected tag is opened for the first time, the AI system will record it, preventing fraudsters from copying it. This way, if anyone tries to copy the product, they’ll get caught, and the product won’t be considered valid. The team filed the patent in the US and India and has plans to do the same in Europe as well. They’re aware that approving their patent will probably take several years, but once that happens, their innovation could help many brands in their fight against counterfeiting.
AI tool shows gender bias
But AI isn’t always the right solution, and like with any other tech, there are limits to what it can do. The e-commerce giant Amazon realized this after its AI recruitment tool showed signs of discrimination against women. Amazon worked several years on the project, hoping to develop the perfect tool that would rate job candidates and recommend only the best ones for a specific position. However, things don’t always go as planned. Amazon’s researchers were surprised to see their AI solution was identifying only male candidates as potential employees, while completely neglecting female job applicants.
The issue lies in the data used to feed the system. This algorithmic system was trained to rate candidates based on previous resumes sent to Amazon, which were predominantly male. Although researchers tried to make the system less biased, they failed, and the solution even identified all-women’s colleges as “less preferable”. Moreover, the same system recommended candidates without the required skills for a certain job position. All this encouraged Amazon to abandon its project, and the company claims it never used it to make real decisions about recruiting new employees.
The future of AI
Considered as “one of the major forces for society and lifestyle of the next decade”, AI provides businesses with tremendous opportunities – hence all the buzz around AI outsmarting us and making us obsolete. An AI takeover still remains a hypothetical scenario, though, but what’s certain is that AI will have a disruptive effect on many industries. We have evidence of this already. Financial institutions are harnessing the power of chatbots to improve customer experience and simplify banking. Furthermore, AI could help companies win the fight against counterfeiting and protect brands’ profit and reputation. This is just a small portion of what AI can do, and as this technology continues to progress, we can expect even more innovative applications in the future.