by Sakshi Dewan
Digital Strategy Director, PHD Global Business
AI was the buzzword in 2017 and it will continue to be under the spotlight in 2018.
There isn’t a conference, talk or article where the words artificial intelligence, machine learning and deep learning aren’t used with great enthusiasm. This excitement has transcended the world of tech and marketing and AI has captured everyone’s imagination. Successful TV shows like Humans and Westworld have given us a glimpse into a future that has tantalising possibilities but is also worryingly dystopian.
So will robots be taking over in 2018? Not likely. While we are making progress in the field of machine learning there are significant barriers that need to be scaled. The big one is that we need greater advances in quantum computing to build neural networks that can truly fulfil the AI promise. We are not there yet, but 2018 will see significant developments in quantum computing.
Another challenge is to guard against AI hacking. Digitalisation has brought huge benefits to our society but has also made our modern systems vulnerable to cyber-attacks. Hacking is a big concern for AI too. As we build powerful applications with wide ranging implications for human society, security of neural networks will need to be a top priority. 2018 will see a focus on security as well.
But the biggest challenge in my view is that of perpetuating human bias through machine learning. AI helps machines become smarter through a continuous process of ingesting data and learning from it. The issue is that developers can often unknowingly create bias in the learning mechanisms through feeding machines limited data or by passing on their own subjective bias while coding. This is a hard one to address, but the first step is for developers to recognise this issue and seek diverse inputs for machines to learn from. In the early days of AI development this should be a big priority, but in the rush to generate applications, it tends to be deprioritised. This could have far reaching consequences in the future, so this is more of a wish than a prediction for 2018!
In spite of these barriers, 2018 will be a promising year for some interesting developments in AI.
Training AI to understand the nuance of language will be a big focus this year. Machines are better than ever at working with text and language. Facebook can read out descriptions of images for visually impaired people. Google makes suggestions on how you should reply to your emails. But they don’t always get it right and can still be a bit clunky.
Human language is incredibly complex, with a lot of nuance and reliance on context. “We’re able to take concepts we’ve learned and combine them in different ways, and apply them in new situations,” says Melanie Mitchell, a professor at Portland State University. “These AI and machine learning systems are not.” This rings so true when I watch my two-year-old point out situations as funny or call herself a “cheeky monkey” with the timing of a seasoned comedian. Alexa is no match for her!
Tech companies are investing heavily in this area and looking at different approaches to teach machines the meaning of words. Facebook researchers are trying to teach software to understand reality by watching video, for example. Google has been developing software that tries to learn metaphors. Deep learning has already made Google Translate far better at understanding and translating languages, cutting down error rates by 60%.
These advances in the AI’s ability to learn languages will mean a direct improvement in voice enabled interaction with machines. AI powered voice assistants have been the biggest consumer tech trend in 2017 with everyone jumping into the fray. Amazon’s incredibly successful Echo, Google’s Home and Apple’s Homepod all vying for what is expected to be a big growth segment.
Most of the interactions with these devices at the moment are pretty basic. We interact in a simple question and answer format for very functional tasks – play music, “will it rain?” and so on. But as machines get a better understanding of our language, these devices will get much smarter and the range of tasks that they will be able to assist us with will increase tremendously. From functional assistants they might turn into social companions.
Chatbots are another application that will get a lot more sophisticated and frictionless as AI develops its understanding of human language. Chatbots’ capabilities will be enhanced from very functional to a lot more developed as they will not be limited to templated responses. This will open up further opportunities for businesses.
But what does all this mean for us as marketing communication experts? I think there are two great opportunities here. The first, is an obvious one; harness new technology to deliver delightful experiences for people. Advances in AI will help us delight our customers through communication that is relevant to their context and needs in a way that just hasn’t been possible until now. If our voice assistants and chatbots begin to understand the nuance of language, we move away from having transactional interactions to personal conversations with these machines. They in turn learn to understand people’s immediate context and emotional need states. This means that we will then be able to communicate with people not just based on their age, gender, search and browsing history but their actual emotional need state. That sounds like a very promising opportunity that will require us to rethink our messaging approach all over again.
The other opportunity is a less obvious one, but equally exciting. In the future, as our machines become smarter and potentially make a lot of decisions for our consumers, we will end up marketing to machines as well as people. And machines are different from people, so we will need to develop an entirely new toolkit to communicate with them – think appealing to algorithms rather than need states. This will be a very interesting challenge for us.
AI does hold a lot pf promise in the field of marketing communication and far beyond. Ultimately, AI will be what we make of it, so we need to avoid the pitfalls and mistakes of the past. Only by preparing for the future will we succeed in it. Let 2018 be the year of AI!
This article formed part of ‘PHD Perspectives’, click this link to read the full publication http://bit.ly/phdperspectives