Dec 29, 2016 | Stuart Frankel
2017 Predictions for Artificial Intelligence and Communication
Artificial intelligence (AI) has officially gone mainstream. Industry expert, Gartner, named AI as its number one strategic technology for a second year in a row. The acquisitions race among giants like Google, IBM, Salesforce and Apple to purchase private AI companies keeps heating up - 2016 alone saw 40 AI-related acquisitions and our own research found that 62% of large enterprises will be using AI-technologies by 2018.
Since everyone seems to be talking about AI broadly, we focused our predictions this year on what we see happening with communications and AI. As a leader in this area, Narrative Science is working with enterprises to close the communication gap between man and machine. For 2017, our predictions are related to how we’ll communicate with computers and other devices, how AI systems will communicate with each other, and how we’ll communicate with each other about AI.
Before we dive into 2017, let's review our predictions from last year to learn what proved true and where we were a bit too early with our predictions.
- New AI inventions will explode - Nailed it! - The rise of open-sourcing AI platforms lowered the barrier for entry for developers, and 2016 saw the rise of new AI startup companies.
- BI platforms will enter a new era - Nailed it! - Gartner calls it “Modern BI” but no matter the label, BI platforms entered a new era as they integrated intelligent features like dynamic narrative generation and NLP that transform how employees interact with and understand their data.
- CEOs will demand transparency from intelligent systems - Not quite yet. - This is a growing trend as evidenced by occurrences like the EU establishing the General Data Protection Directive, which goes into effect in 2018, and states that individuals subject to automated decisions by intelligent systems have a ‘right to explanation.’
- (Data) size doesn’t matter - Nailed it! - The Big Data hype has subsided, and it's no longer just about the size of your data, its what you do with it that matters.
- Portions of data science will be automated - Not quite yet. - 2016 was more about defining the role of a data scientist, identifying skillsets needed, and the launch of new training programs. While new technologies emerged to make analytics software easier to use, it isn’t enough by a long shot - IDC predicts a need for 180K people with deep analytical skills in the US by 2018 and a requirement for five times that number of positions with data management and interpretation capabilities.
- Innovation teams in large companies will increasingly become a competitive asset - Not quite yet. - While the latest research shows that 131 corporations run their own accelerators or innovation teams, there is little evidence that these innovation teams are a competitive asset. Instead, they are the first step in a longer journey. The Innovation team may pilot an exciting new technology but if they can’t influence internal adoption, the technology never takes off and the innovation is stopped in its tracks. Companies who infuse innovation into their culture by actively fostering it, rewarding it, and frankly, expecting it, will win the innovation race. Our customer, USAA, is a great real-life example of this practice in motion.
AI & Communications Predictions for 2017
Now that we've taken a look at how we did in 2016, here are our five predictions for Artificial Intelligence and Communications in 2017:
Prediction 1 - The movement towards conversational interfaces will accelerate
The recent, combined efforts of a number of innovative tech giants point to a coming year when interacting with technology through conversation becomes the norm. Are conversational interfaces really a big deal? They’re game-changing. Since the advent of computers, we have been forced to speak the language of computers in order to communicate with them and now we’re teaching them to communicate in our language.
Search engines like Google and Bing have already made big moves enabling search queries via spoken word while Facebook launched an AI-effort, DeepText, to understand individual users’ conversational patterns and interests. Meanwhile, the move toward natural language interfaces has already picked up steam with the explosion of companies focused on enabling chatbots, digital assistants and even messaging apps eclipsing social networks in monthly activity.
Beyond 2017, think of a future when we can casually ask our personal devices for information regardless of subject - “How much money do I have in checking?”, “When was my last physical?” or “What restaurant within a 10-minute driving distance has an open table for 2 people?”
Prediction 2 - Design will begin to evolve to increase our trust in AI
If people don’t trust AI, they won’t use it. In the next year, designers will begin to apply knowledge of human interaction, specifically in the area of how we earn trust and respect, to AI systems.
Elements of communication like tone, sentiment, timing, visual cues and word choice combined with AI technologies like natural language generation that increase transparency into how these systems operate will play a role in helping users trust and rely on AI systems.
Stanford’s recent study on AI’s impact over the next 100 years states it well:
“Design strategies that enhance the ability of humans to understand AI systems and decisions (such as explicitly explaining those decisions), and to participate in their use, may help build trust and prevent drastic failures, it's critical that engineers and designers create systems that communicate freely about how they work.”
In other words, if my AI-powered home monitoring system unlocks my home for an unscheduled visitor in the middle of the day, it better be able to explain why.
Prediction 3 - We’ll start talking about how AI systems talk to each other
In the next year, efforts will begin to create universal standards for AI to AI interactions. Without standards, AI technologies will increasingly become siloed or worse, interfere negatively with each other when multiple AI systems are involved in determining a single outcome.
Imagine driverless cars on a collision course without the means to communicate with each other or an enterprise with multiple siloed AI systems that has a predictive analytics system moderating decisions about production levels but another AI system with a different data source that has indicated production needs to change.
2017 will be the beginning of talks among the tech giants, relevant industry associations and governmental bodies to establish universal AI standards.
Prediction 4 - Discussions of imbedded bias in AI will grow
In 2016, examples that reflected the multiple sources of bias that can occur within AI systems. Some of these sources include the data used to train systems, users’ interactions with the systems, similarity bias and the bias of conflicting goals.
Most of this bias currently goes unnoticed but as AI usage grows and increasingly impacts people’s lives, recommendations need to be established for acknowledging and addressing systems’ biases or AI will take a major hit impeding future progress.
Prediction 5 - Enterprises will start to demand ROI from their AI
Companies will begin looking for demonstrable value and ROI proof points from AI technologies. While funding for AI-related startups keeps increasing - in the last 5 years alone, investments in AI have grown tenfold from $94M in 2011 to $1049M in 2016 - we’ve seen few real commercial applications surface. Most often these technologies are piloted by Innovation teams or an R&D department. 2017 will be the tipping point when companies start questioning their investments and AI will have to grow up.
Its pretty amazing to think that just two years ago we were talking about AI and robots coming to kill us. Tech luminaries were proclaiming AI would bring upon the apocalypse and now, some of these same people are founding organizations to push AI to its’ limits. So much has been accomplished in a short span of time, and we’re now starting to realize the benefits of partnering with AI versus fearing it.
I’m already looking forward to next year when I can review my predictions to see how we fared - or more likely, I’ll be asking my intelligent system to tell me about my hits and misses. Hopefully, I’ll do well.