By José de la Rubia, Policy Analyst, OECD
This article is by Featured Blogger José de la Rubia from his LinkedIn page. Republished with the author’s permission.
A recent article by James Titcomb in The Telegraph on Professor Stephen Hawking last speech at the opening of the Leverhulme Centre for the Future of Intelligence brought to my mind the concerns and the promising hopes on advances in AI.
Artificial intelligence has been around since the start of computing and has had many false starts. The reality did not live up to the expectations set by science fiction. Accordingly, for many years, the majority of people’s understanding of AI was confined to university laboratories, corporate skunk works, research parks, and that movie with Haley Joel Osment and Jude Law. Attempts to introduce AI products and services into the marketplace and for the broader benefits of society were unlucky. Computing power was insufficient, and the profusion of structured data — let alone a knowledge of what to do with said data — was not yet upon us.
AI has been on the cusp of the mainstream for the past 40 years, but 2016 is the year it has become a buzzword — incorporating machine learning, natural language processing, voice recognition, and data mining, to name a few technologies. Major corporations are now striving to integrate AI into their products.
A new machine learning program from Google wrote its own piano song. Advances in natural language processing are developing at such a quick pace that scheduling a meeting most likely involves cc’ing your personal AI assistant. IBM Watson, an AI-powered computer, is already able to deliver life-saving medical diagnoses in less than 10 minutes. Soon, students from research universities will be training it to fight hackers. According to data from CB Insights, over 200 AI-focused start-ups received venture capital funding in the first two quarters of 2016 alone, underscoring the massive amounts of time, resources, and capital being funnelled into the space.
Why AI, and why now?
I grasp three core developments in recent years that have brought on this shift:
- Increasing layers of data that used to be analogue are now digitized, so all of a sudden there are massive data sets around people and systems, how they’re behaving, and what they’re doing.
- Processors have become faster, and the processing architecture has become far more sophisticated – massive amounts of data can today be analysed in real time, which can inform immediate action.
- Significant progress has been made in the advancement of the underlying science and ultimate capabilities enabled by AI’s innumerable applications.
Because AI has the potential to affect so many industries and in so many ways, we think of it as a layer across multiple sectors, rather than a sector in and of itself. Look to the early 2000s and observe how the mobile inception has touched virtually everything we do today.
We envisage similar results with artificial intelligence in the years to come. But how will that happen?
Accelerating the commercialization of AI
As AI progresses, so too are the resources and expertise that are becoming available to entrepreneurs and start-ups try to commercialize the technology. Transformative partnerships are growing up, such as those between universities and technology/venture industry experts, helping to introduce usable AI at a large scale.
From chatbots to data analysis and automation, AI start-ups are developing solutions that make existing processes and businesses more efficient. Look at Salesforce, which, despite making certain aspects of the traditional CRM process easier, is more or less still an analogue operation. New AI applications, including machine learning and natural language processing, makes the process more efficient by providing a layer between data and the human interface. As we move toward a text message-based society, this efficiency becomes vital for AI business cases.
Chatbots that replace face-to-face interactions in brick-and-mortar stores have become staples of ecommerce websites and have produced great results. Talking with a person converts product page views to sales. Yet a person on the receiving end of a chat is limited to the number of conversations a human can have — and thus the number of customers the company representative can possibly give service at any one time.
Herein lays one of the many opportunities for AI and AI start-ups. The more that conversations move to a medium that doesn’t require audible interaction, the more that AI can be used as a layer to supplement or even replace basic conversations.
The rapid increase in chatbot usage and investors’ continued foray into pouring capital into AI bot development attest the fact that people are now ready to interact through text-based messaging. Wade & Wendy, an emerging AI company, is an example of how two different AI personalities will soon be able to integrate with job candidates and recruiters, respectively, in order to make the process of recruiting and career advancement more efficient, transparent, and ultimately more human. This layer of technology gives us a window into what’s possible when a machine can take over the repetitive but vital task of servicing potential customers as commerce moves to web and mobile.
While AI-powered chatbots – those that use responses that can refer back to context throughout a conversation – are still a few years away, those that provide rule-based responses are already critical assets for businesses large and small. As with the semiconductor industry in the past and the autonomous car industry today, universities are working more closely than ever with industry partners to make AI as transformative as the first silicon wafers were in the 1950s and as impactful as the birth of the iPhone in 2007. And that’s an exciting prospect – with all due respect to the synthesized voice of the Oxford genius.
Now what worries me about AI chabots is if they follow the path of human development, there will be a few awkward years when machines become teenagers…
Originally published on LinkedIn