A year of AI events points to AI predictions for 2020

Looking back at last year's predictions

Last year, we made a number of predictions about how AI would play out in 2019. While some of these predictions were a bit hit or miss, it wasn't entirely a washout on how we saw the year playing out. Last year we expected that AI assistants -- especially voice assistants -- would see greater play in the enterprise. While Amazon did announce the release of their Alexa for Business platform expanding the reach and use case for enterprise voice assistants, by and large the assistants stayed out of the enterprise market. In fact, Microsoft announced that it would actually pull back from Cortana this year, keeping it a part of their desktop experience but limiting expansion otherwise. Google continues to struggle with Google Home adoption, and Apple's Siri product hasn't garnered widespread enterprise adoption. It remains to be seen whether voice assistants will become a fixture in the enterprise environment.

We predicted last year that 2019 would be the year of the pre-trained model and improved third party data sets. We're going to double down on this prediction for 2020 and say that 2020 will be the year of model-as-a-service. This means that we'll see an increasing array of companies and organizations making their models available for use, shifting the market from being mostly focused on producing machine learning models to consuming them. This will necessitate the growth of "machine learning ops" as the market starts to shift towards consumption of third-party models as the early majority of customers enter the market for AI.

We also predicted that chatbots would dominate customer service this year. While there wasn't any groundbreaking news on this front announced in 2019, it's clear that the trend towards self-service and AI-enabled conversational apps continues unabated. Customers continue to show their preference for automated, machine-mediated conversational engagement and certainly nothing happened in 2019 to slow this trend down.

Our final prediction was that 2019 would finally be the year that responsible and ethical AI would come to the fore, given the pressure on companies like Facebook, Amazon, Google, Microsoft and others now that customers are demanding more visibility and control over their data. In 2019, we saw several jurisdictions banning the use of facial recognition technology and even a movement to declare how and where cameras with facial recognition technology will be used. In our AI predictions for 2020, we're seeing a variety of efforts to formalize the declaration of transparent and ethical AI, something we'll discuss later in this article.

Governments accelerate their strategic push on AI

2019 was also the year that governments doubled down on their AI investments. The Trump administration updated their national artificial intelligence strategy and set a $973M non-defense AI R&D Budget for FY2020, which started on October 1, 2019. Not to be outdone, China took the top spot in the ImageNet challenge for facial recognition, thereby ending the challenge altogether.

However, the US and China aren't the only ones competing in the race for global AI domination. Additional countries have made strategic investments, strategies and promises around AI as well. Australia dedicated $29.9 million in the country's annual budget to promote and guide the development of AI. Estonia released an AI strategy this year which focuses on accelerating AI in private and public sectors throughout the country and is developing a legal framework for the use of AI in the country, including upcoming legislation on AI liability. Similarly, Lithuania released The Lithuanian Artificial Intelligence Strategy in April 2019. While much of the attention has been on the large world economic powers, smaller states have much to say about AI's future.

Growing questions on data usage

In 2019, it was clear that general consumer sentiment on corporate data usage had turned negative. As people become increasingly more aware of their data footprint, they are expressing concern about how companies are making use of this data. In response, governments have passed legislation restricting the usage of such data. GDPR has been in effect for over a year and companies need to comply with the various rules and regulations of this and other governments' legislations. In 2019, California passed the California Consumer Privacy Act (CCPA) that officially takes effect in January 2020. This law, combined with GDPR and enacted in the heart of Silicon Valley, will no doubt shake up the world of data, which has dramatic impacts and consequences on the AI industry.

To further these goals and assist government organizations, technology vendors and enterprises, as well as organizations such as the Advanced Technology Academic Research Center (ATARC), are coming up with standards-based efforts such as a transparency score that will be used in conjunction with emerging model-as-a-service and machine learning ops efforts to provide more visibility into how models are created, collected and modified.

2019 AI newsmakers

AI continues to make news in 2019. On the venture capital front, the robotic process automation (RPA) market is funneling hundreds of millions into software and workflow automation tools that many see as a "gateway" to AI. However, we're already starting to see cracks in the RPA market. Microsoft entered full force this year with their UI Flows RPA capability in their newly rebranded Power Automate platform. This will pose a significant challenge to the RPA incumbents as Microsoft is already claiming over 150,000 customers in this industry. Almost on cue, after their multi-million-dollar event in Las Vegas, UiPath announced a round of massive layoffs, attributing those layoffs to unsustainable growth as it enters a period of stabilization. Despite signs of overheating in the RPA market that might be entering a phase of consolidation, Automation Anywhere announced a very large Series B round at the end of the year also showing that investor appetite in RPA hasn't diminished in the least.

In other major AI news, OpenAI released the full, unbridled version of their GPT-2 model, which many feared would be put into the hands of bad actors. Citing no evidence of misuse so far, we'll have to remain cautious about the use of these big models. It is possible that 2020 will be the year we see these powerful models applied to negative use cases, especially during what will no doubt be another contentious electoral year in the United States.

In addition to the technology side, 2019 was the year in which it became very evident that humans are still needed to make AI work at scale. Data labeling and preparation is a hot area of investment and attention as companies realize that humans are needed to annotate and label data to feed supervised machine learning systems. Without these humans, much of what has been produced to date would simply not be possible. As such, in our AI predictions for 2020, we will no doubt see the continued growth of the human-powered data labeling business.

The AI momentum continues at enterprises

AI continues to be a driving force in many different industries and market sectors. As we covered in the past year of articles, AI is being increasingly seen in a wider array of organizations. We profiled how AI is making its way into the IT organization from IT service management to project management to IT operations and network administration. We're also seeing AI in a wider array of industries including mining, construction, fitness and wellness, legal, finance, accounting, advertising, sales, consumer packaged goods, transportation, knowledge management and management of professional services organizations. It seems that AI is present in every corner of the market.

We saw the narrow application of AI being applied to recognition, conversational patterns and autonomous systems. We also saw the increasing use of AI for cognitive automation, AI for patterns and anomaly detection as well as predictive analytics, goal-driven systems, and the increasing use of machine learning and AI for hyper-personalization.

We're also seeing some interesting and creative uses of AI, from AI that is being used for music and art generation to other ways in which AI is helping enhance the human experience. There's now increasing use of collaborative robots (cobots) in retail as well as greater use of AI to assist with personal and business relationships. AI is helping solve some of our world's most pressing environmental problems, but possibly also adding to some of the challenges of climate change.

In sum, it seems that 2019 is a continuation of what we have seen in recent years. There haven't been any showstoppers that have prevented AI's continued use and growth, and there's no indication of a pullback in interest or investment at the enterprise, venture capital or country level. The state of AI in 2019 is strong and it sure seems that it will be even stronger in 2020.