AI Now & Its Future

The CEO and co-founder of Sangwa Solutions, Chris Earle, defines AI as simply, “a machine that can learn. Generally, people have this notion that there is something called ‘The AI’ which would be some sort of singularity of AI like, the AI ‘overlords.’ But I think of it more as a facet of technology.” We have a general understanding of the start of AI, but what does it look like now? What does the future look like? Artificial intelligence is seen in every part of our lives.

Art

Artificial intelligence art is a style of art that uses software to produce work. Instead of paint, Artist Refik Anadol uses machine learning to create dynamic pieces of work. His ‘data sculpture’ titled “Machine Hallucination” uses an algorithm to compile 10 million images of New York City and poetically present them. The algorithm is used as a narrator to show the entire lifecycle of the city. Anadol explains that everyone experiences the city through a different lens, but his art is unbiased. In the exhibit, he describes the piece as looking like a machine is dreaming. The ‘sculptures’ are projected onto the floors, ceilings, and walls of the room – showing anything can become an interface. All of his work questions “can data become a pigment?” Watch the full experience here.

Music

AI has reshaped the way we make and produce music. The website DeepBeat allows anyone to use machine learning to create a synthesized rap. DeepBeat generates lines using the RankSVM algorithm and a deep neural network model – while still capturing the creativity and meaningful words required. DeepBeat uses a database of existing songs to find the most suitable and relevant series of lines. According to an evaluation of the app, “DeepBeat outperforms the top human rappers by 21% in terms of length and frequency of the rhymes in the produced lyrics.” Objectively speaking, the website improves efficiency when writing lyrics.

Toys

AI is seen almost anywhere, even in children’s toys. In 2015 Mattel released ‘Hello Barbie’ which allows children to have interactive conversations with Barbie. This software uses machine learning by evaluating the conversations to produce more meaningful ones in the future. Barbie uses voice recognition to ask questions about family, friends, and school. Through the mobile app, kids can play games with Barbie and listen to previous conversations. In 2018 Mattel released the ‘Hello Dream house’, a dollhouse that responds to commands such as “open the door” or “let’s have dinner.” As a society, we are using AI at a young age in our everyday lives.

Health

Earle speculates that AI has been most impactful in the healthcare industry. Woebot Health is a chatbot app that is trained to assist in Cognitive Behavioural Therapy (CBT). Just like ‘Hello Barbie’, individuals answer questions posed by the Woebot and build a personal connection with the technology. In comparison to conventional therapy, Woebot is much more accessible – but are chatbot apps as effective? Many people are using the app, “in early 2018 [Woebot] sent 2 million messages per week to users in 135 countries across the globe with an estimated few hundred thousand monthly active users.” It is hard to say if this new technology is efficient, but the concept is interesting. 

The app Maple allows you to connect with doctors at any time or place. It is a form of virtual healthcare that matches you to a physician based on the symptoms you provide the app with and your location. The doctor can then provide you with advice, treatment, or even a diagnosis. This is extremely beneficial to areas with a lack of healthcare professionals or those with accessibility issues.

Grocery

Amazon created AmazonGo using computer vision, deep learning algorithms, and sensor fusion grocery stores with no lines or checkouts. Amazon calls it, “just walk out technology.” You simply pick the items you want and leave, reducing any interaction with employees, and your amazon account is charged. This makes grocery shopping faster and easier. RBC analyst Mark Mahaney describes the experience to be “almost like shoplifting, except it’s legal.” Not only is it legal, but it creates an efficient and pleasant shopping experience. Due to its high level of interest, Amazon reported opening 3,000 AmazonGo stores by 2021. The majority of consumers are giving raving reviews – one said they were “In & out [of the store] in 27 seconds flat with breakfast!” This technology is the future of AI and the future of retail.

Limitations

AI works well for repetitive and predictable tasks. For example, jobs in the service industry can have extremely repetitive tasks that, with the right solution, could be automated. However, soft skills cannot be replaced by machines. Interactivity between consumers and staff is a key responsibility in certain jobs. However, The connectivity between people is more important than the product or service.

Going to the barber or hair salon is much more than just getting a haircut. When you sit in the chair, most likely, you engage in meaningful conversation with the stylist. Any career that includes high-level interaction with their clients is limited to the amount they could be automated. Even routine jobs like analyzing x-rays can be automated. Given the proper background knowledge, a machine could do this job at a lower cost.

Although the majority of things could be replaced by AI-based software, there is one for sure that cannot – nurses. Not only is this a complex job, but it requires a high level of emotional awareness. Making a mobile robot that has the abilities to perform health care duties would be extremely elaborate and we are not at that level yet. Furthermore, nurses have compassion which is intangible for robots to have. “That doesn’t stop us from trying” – although we have not found a way to make a robotic nurse, there is still the opportunity to in the future.

The Future

Earle explains that the incentives to automate a job relies on the number of benefits that it will produce. Considering the cost, the complexity, and the repetitiveness of the job. Is there a reduced need for human input? Yes. Chris begs the question, “how many blacksmiths do you know?” We are in a transition period where old jobs are being replaced, but there is a creation of new ones. To mitigate against the need and desire for humans, it is key to have a set of skills that is not replaceable and in demand. Simple things, like knowing how to use Microsoft Word and Excel, are essential these days but were non-existent a few decades ago. As time continues, these job requirements will advance and we will have to adapt to the digitized future. Read about more emerging technology here.