Artificial Intelligence (AI) is no longer the stuff of sci-fi. It’s in our homes and ordering our dollhouses. But what’s next?
2016 may be looked back on as a year with unexpected outcomes, but it was a brilliant year for those who work in artificial intelligence. The proliferation of AI devices has brought us one step closer to the IoT dream of self-managing homes, and as Google and Tesla fight to bring the first widespread rollout of self-driving cars, the battle for greatest AI of them all will only intensify. So, if 2016 was the year that AI-at-home became a reality, what does 2017 hold?
1. Competition Will Step Up
Early adopters have been testing AI devices for a long time, but now that there’s a proven appetite for AI solutions among the general public, many companies will seek to elbow their way into the market. Companies that are already leaders in the space - Google, Amazon, Microsoft, IBM and Salesforce - will boost their internal AI innovation. A trend well illustrated by the fact that IBM decided to invest more than $1 billion in their Watson technology. There also stands to be huge investment into training data, given that even the best machine learning (ML) solution is useless without any data to learn from. As leading companies push to get the machine learning market edge, we’re likely to see a wider array of companies trying to follow their lead.
2. But More ≠ Better
As more companies try to ride the wave, we are likely to see a significant growth in the number of businesses attempting to create their own version of an AI. While competition is healthy, it’s only good if the technology works and building a good AI is a very delicate art. On the one hand, the potential here for bad tech that will devalue the overall good name of AI is high. On the other hand, this is also good news for companies who have built AI technology worth its name as they will truly distinguish themselves from the crowd and from the competition. The potential of AI will nevertheless drive companies to keep trying their luck in the field and companies that have been working for years to perfect intelligence algorithms will now face competition from algorithms that are simply good enough.
3. Focus on Business Rather Than Commercial
As mentioned above, the market leaders in the AI sphere will pour resources into developing strong business-focused solutions. In 2017 we’ll start to hear more about B2B AI solutions that focus on streamlining internal business processes, and software solutions that go far beyond the nice-to-have commercial AI’s like Siri, Alexa, etc. Machine learning solutions will be focused on adding true value, not only in our homes, but also for large corporations and businesses. AI solutions for structuring data and documents will for example be particularly valuable for large businesses. In a time where 2.5 quintillion bytes of data is produced every day, it becomes crucial for large corporations to turn their unstructured data and documents into insightful and actionable structured data. LEVERTON is already applying deep learning technologies to shed light on enormous amounts of unstructured documents for global corporations.
4. Data Scientists Will Finally Get to Focus on ML
There are a lot of frustrated data scientists in the world. Having originally been hired into machine learning positions, many find their days dominated by the collection and cleaning of data, and the day-to-day maintenance of backend software. Data scientists rely on high-quality data to be able to perform their jobs well, but just having that data to hand takes a lot of work. Often, data scientists are forced to either ask developers to spend valuable time cleaning and categorising, to take a risk on outsourced resources, or to simply do it themselves. This year, services like Amazon’s Mechanical Turk, a platform that helps developers complete routine tasks which require human intelligence, will save data scientists a lot of costly time, and allow them to apply the necessary machine learning solutions at scale.
5. Expect A Lot of Ethical Discussions
Ethicists have long been discussing the potential moral quandaries of artificial intelligence. TV shows like Westworld and Black Mirror have taken some of these debates to their most extreme conclusions, but all they serve to underline is that people are nervous of the potential implications of AI, and are not at all sure how to move forward with it. How can an algorithm be taught to make the right ethical decision? Should, for example, a driverless car risk the life of the passenger to save a pedestrian? Should an AI be able to make decisions in life-or-death scenarios? If so, who’s at fault when they go wrong? This debate will rage on for some time yet, but 2017 will likely be the year that it becomes a hot topic of public debate.