Future Predictions of AI

2017 was official “The Year of AI”, with publications like The Wall Street Journal and Forbes making that call. And the venerable Time magazine asked the seminal question in our minds, “Are Computers Already Smarter than Humans”? AI outperformed professional players, most notably in beating a Chinese master at “Go”. Real-time speech translation is becoming more powerful every day and in next five years, my Spanish would be as good as my Japanese, thanks to the AI and machine learning. These are just a few of the milestones AI reached in 2017. So what’s next? As artificial intelligence continues to disrupt multiple industries — I think 2018 is going to be transformative. Here are my top predictions:

The Cambrian Explosion of AI Technologies in 2018

In a recent survey, McKinsey estimates ~40% of leading companies have deployed AI technologies or machine learning models in their business. While an all-encompassing, multi-functional AI that assists us in every aspect of our life is many years away, we see the “Cambrian Explosion” of different AI technologies and applications, each with a specific niche.

This is indeed a good sign. Just like the mass diversification of complex organisms over a relatively short period of time during the early Cambrian age filled every niche 541 million years back, various AI technologies and experimentation within different industries will make AI more mainstream. A couple of areas of personal interest to me are:

  • Software 2.0: AI writing software: Unlike software 1.0, which is written in Python, Java or C++ with explicit instructions, Software 2.0 is written in neural network weights with no humans are involved in writing code. In short, AI writing your code
  • Explainable AI: understanding the black box: Most AI in existence today are “black boxes”; they offer little insights on how the outcome was reached. Explainable AI aims to develop techniques that produce more interpretive models. This is critical in long-term: an explainable and transparent AI is critical to establish trust and encourage faster adoption

AI will Create Contents we’ll Enjoy

Noticed how Netflix is getting better at recommendations of late? Well, not only it gave you better contents, Netflix has reportedly saved over a billion dollar on content creation and management by using AI. On the news front, leading UK news agency Press Association (PA) is betting that AI can fill the gap left by former reporters and shuttered local newspaper offices. Using Natural Language Recognition (NRG) technology, PA aims to create 30,000 localized news reports every month. Muses Jan Kautz, an AI researcher at NVIDIA, “imagine a future music service that doesn’t just play existing songs you might like, but continually generates new songs just for you.” Many contents that you will soon see, hear or read may indeed AI-generated.

Figure 1: Many contents that you will see, hear or read would be generated by AI soon, including some incredible arts being generated by Google’s AI (Source: Google)

Smart Assistants will Run our Lives

Our personal assistants (PAs) are continuously getting smarter. Today we have a plethora of chatbots, recommendation engines and personal assistants, proving bits and pieces of answers and assistance. As these PA bots learn more about our daily routine and integrate that information across platforms and applications, I can imagine a day that I won’t forget about dinner, which I occasionally do when busy. My smart assistant will know which days of the week I typically work late, confirm my location from my phone and order my favourite take-out meal. Other days, with awareness of ingredients in my fridge from my shopping order my smart PA will suggest dishes cooking at home.

Talking to Machines will become Normal

Amazon sold “tens of millions” echo devices this Christmas, yet besides occasionally asking for weather or for a song to play, most of us have not gotten used to the idea of talking to a machine. In fact, a chatbot for customer services, even when it provides faster and more accurate information than a human customer support agent, isn’t preferred by most users yet. The reason is threefold: first, Alexa/ Siri/ Google voice still needs a lot of practice before it can understand us, especially for those who have accents. Friction will reduce as the speech recognition technologies improve.

Second, as natural language processing improves in 2018 and gets better at picking up on human intent and emotion, chatbots and messenger bots will understand us better and provide more contextual answers. Finally, we will have more opportunities to interact with machines. For example, many brands such as Disney, Marvel and Unilever have released chatbots to reach consumers in the channels they use daily. Innovative chatbots like Endurance is helping dementia patients having a better life, while Casper is helping insomniacs get through the night better. In the professional front, MedWhat is making medical diagnoses faster for doctors while Roof. AI helps real-estate marketers to automate interacting with potential leads and lead assignments.

AI, chatbots and conversational technologies are about to disrupt virtually every industry in 2018, so be prepared to see them in your house, work, and beyond! For those who still doubt that 2018 is the year for it, a tweet I saw on the first day of 2018 confirms it!

 – Santanu Bhattacharya

Posted in AI

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s