By Steve Kilpatrick
Co-Founder & Director
Artificial Intelligence & Machine Learning
Last year, we put together our list of top trends in artificial intelligence for 2018. Those AI trends focused on increasing acceptance and understanding of AI in business, consumer, and legal applications. Those trends largely bore out over the course of the year, and the continued development of those artificial intelligence trends is what to look for in 2019.
This year, we’ve put together another annual post of what to look for in artificial intelligence. As AI becomes more advanced and simultaneously more mainstream, we’ll see a proliferation of applications using AI in 2019. This rapid development brings a lot of opportunities, but it also raises concerns about the ethics and uses of this powerful emerging technology.
Trends in Artificial Intelligence to look out for
1) More Jobs, Not Less
Editorials love to proclaim the end of work with the rise of AI. The argument goes that algorithms will be able to perform the knowledge work of many professions, leaving little for human workers to do in the future. However, according to survey research from PwC, the growth in AI will lead to more jobs, not less. 31% of executives PwC surveyed are worried about meeting the demand for AI skills over the next five years.
There’s the obvious need for developers and data scientists to create and train algorithms. In addition, companies need data engineers to clean, structure, and deliver data to the algorithms. Skilled technologists are certainly needed and hard to come by at the moment, but AI will also create less skilled jobs. Humans need to clean and label datasets by hand for certain applications, especially visual and language recognition.
2) Increased Enterprise Use
2018 was a major year for AI development. In 2019, expect machine learning algorithms and neural nets to go into production on a wide scale within companies. Increasingly, companies will use AI to automate and simplify tasks for employees. This is the low-hanging fruit for internal applications of AI, and it could lead to massive productivity gains.
More challenging, but perhaps more valuable, are the second-tier applications of AI for making predictions and strategic decision making based on company data. This is the realm of data science, which is rapidly growing in popularity amongst companies for business intelligence. Applying machine learning and neural networks to company data will become standard practice for all companies within the next few years.
3) Smart devices get even smarter
The connected home industry gets stronger every year, and the applications powering the connected home are getting smarter. Smart home systems can now learn and control the lights, security, entertainment, and temperature in your home. Home hubs and smart speakers are becoming the new must-have consumer technology, bringing AI voice assistants, like Alexa and the Google Assistant, into your home. Sales of smart home speakers are exploding. Deloitte predicts 250 million smart speakers will have been sold by the end of 2019. Revenues from in-home assistants will reach $7 billion, up 63% from 2018.
The home isn’t the only place where AI is powering consumer products. Many cars now include smart features, lane keeping, and some level of autonomy. The 2019 models of the iPhone will include a third camera for depth of field and sensing for computer vision and augmented reality applications. Increasingly, AI capabilities are becoming standard for consumer devices.
4) Open Source Algorithms
The development ecosystem for machine learning algorithms is on the rise as well. Most of this development is happening in open source repositories where anyone can use the algorithms and suggest changes. In addition, the major platforms for training algorithms are also open source. The upshot is access to machine learning and neural networks is open and democratised. This is hugely important for the creation of ethical, understandable algorithms. It’s promising that development of such powerful technology isn’t happening behind closed doors.
Another major benefit is the interchangeability and fast time to deployment when using standard algorithms and platforms. This quick ramp-up time will lead to a proliferation of AI in both business and consumer applications.
5) AI Infrastructure is Key
With the rapid rise of AI, it’s tempting for companies to think they can scale up quickly. However, when it comes to deploying AI applications, companies need to walk before they can run. According to research from Gartner, 85% of big data projects fail. To avoid such failures, companies should focus on implementing data infrastructure, practicing statistical fundamentals, and getting easy wins early on with machine learning.
When putting AI applications into production, DevOps is becoming increasingly important to scale highly available apps. Over the course of 2018, Kubernetes-managed containers have become the standard for deploying AI to production. That trend will only continue in 2019 as Kubernetes establishes itself as the dominant container orchestration solution.
6) Trust in AI abilities
In the 1990s, Google was an interesting tool, but few knew how web crawling worked and the users didn’t instinctively trust search results. If you needed to answer a question, Google was a resource, but you didn’t trust it to always be correct. Today, however, we instinctively google even the smallest questions and trust the results.
A similar trend is unfolding for AI. Just a few years ago, talking to the average person about AI would have ended in fears of sentient robots and vengeful algorithms. Today, the public and the media are much more comfortable with AI. As AI gets better at responding to our needs, we’ll trust it to handle more tasks in our lives. As part of that transition, consumer, in-home AI will play a huge role in making algorithms seem trustworthy.
With more people trusting AI to run various aspects of their lives, the ethics of AI become important on a societal level, as does government regulation to make sure AI doesn’t overstep its bounds.
7. Governmental Intervention
Despite the exciting trends, there are serious concerns about the development of AI that we should be worried about. Governments have already started regulating the ways companies can acquire, store, and use customer data. New regulations are needed in emerging areas of AI like facial recognition or voice and image fakery where AI can be used for nefarious or fraudulent purposes. Additionally, governments will need to consider how to prevent bias in AI as AI starts to infiltrate every aspect of our lives.
The good news is governments have started taking notice. Mentions of AI in British and Canadian Parliaments and US Congress have spiked since 2016. Whilst no government has emerged as the leader in AI regulations, 2019 could be the year when we start to see governments taking a strong stance on legal, ethical AI development.
The Future of AI
AI will profoundly change society. It will impact our homes, work, cities, and nations as we integrate AI into nearly every aspect of our lives. We are proud to be working with some leading brands and startups who are on the cutting edge of creating or integrating Artificial Intelligence into their businesses models. We look forward to them and their customers continuing to see the benefits AI will bring in the years ahead.
2019 could be a critical year for the trajectory of AI. The decisions consumers, companies, and governments make this year will set the course for artificial intelligence’s adoption and uses over the decades to come. Expect these AI trends to continue well into the future.
If your company is investing in artificial intelligence and is looking to hire, please get in touch with me. I would love to learn more about the projects you are working on.
If you have enjoyed this article, please consider signing up to receive our monthly newsletter below which includes our latest articles, jobs and industry news.