Trends for AI in 2020 and beyond

Whether it’s machine-written news articles, AI-led cybersecurity or key developments in emotion detection, 2019 certainly brought cutting-edge achievements to the world of artificial intelligence (AI). Looking ahead to the 2020s, what can we expect? AI will certainly continue to develop in promising ways for businesses and consumers alike. The hype that has surrounded AI until now is quickly becoming replaced by a more tangible reality, providing a chance for organizations to reap the rewards of AI in outcome-driven ways.

Here’s how AI will help business leaders and decision-makers push the boundaries of their enterprises to deliver improved business performance: 

AI: no longer just a buzzword. Today, AI is no stranger to most workplaces. According to research from Forrester, 53% of global data and analytics decision-makers say they have implemented, are in the process of implementing or are expanding or upgrading their implementation of some form of AI.

In 2019, we saw a shift in conversations around AI. Organizations started focusing on business value gained from AI, and vendors educated customers on AI maturity. At the same time, companies began looking for roles beyond data scientists to support them in their AI journey.

This year, we will see a continued focus on tackling the “real” challenges with AI, and less hype on becoming vaguely “AI-enabled.” There will be more talk around where you can get started with AI and what team skills you need to apply AI to business. You will hear more on how to achieve business ROI through adoption.

More businesses will realize that putting the right people in charge of their AI program will produce stronger benefits. For example, companies with a dedicated chief data officer (CDO) or chief analytics officer (CAO) are already about 1.5 times more likely to use AI, machine learning (ML) and/or deep learning for their insight initiatives than those without CDOs. By putting more emphasis on AI within the workplace, employees in every department and role will take notice of AI’s increased importance to the enterprise.

Data science teams will continue to realize the benefit of welcoming members with cross-discipline experience, which will further expand the number of non-tech people playing a bigger role in the use of AI – propelled by easier access to ML toolkits and AutoML capabilities.

As AI adoption grows and more and more employees actually experience how AI can augment, improve and even fundamentally transform their strategies and job roles, AI will no longer be relegated to the R&D wing of an enterprise. It will touch virtually every part of a future-forward company. 

Invisible AI implementation. In 2019, we saw great advances in narrow AI, including areas such as satellite imagery, natural language processing (NLP) and computer vision. Going forward, expect to see more breakthroughs in AI research coupled with the commoditization of AI models. AI will be increasingly embedded into domain-driven solutions, allowing it to become invisible yet omnipresent.

The trending use of AI in edge computing is one example. Driven by a more efficient use of resources, information processing and content collection will move closer to the source of the information. AI on the edge also enables lower latency and heightened data privacy. AI-driven inferences, pattern-matching and predictions will blend into the domain applications at the edge, making the user interfaces smarter.

The IDC predicts that by 2022, 75% of enterprises will embed intelligent automation into technology and process development. The same report projects that by 2024, AI will become the new user interface, and 50% of user touches will be augmented by computer vision, natural language, augmented reality and virtual reality. AI will be everywhere, yet never seen. 

Augmented AI: a human-driven future. This year, you will see more visible collaborations of humans with AI. AI will be increasingly designed with humans-in-the-loop, allowing for augmentation of capabilities.

Human-in-the-loop is not just a fail-safe in the event of an emergency; it’s like having a human driver behind the wheel of today’s autonomous cars. However, the role of humans in AI systems goes much deeper than just being on standby for when things go wrong. Human input is necessary to mitigate problems such as AI bias and improve the quality of the everyday decisions of models. Organizations are already paying attention to this need: According to the IDC, over the next four years, we will see 75% of enterprises retraining and developing employees to address new skill sets to optimize human-AI interactions. 

However, with great power comes great responsibility. Forrester warns of three PR disasters that potentially can “rattle reputations” in the tech world this year: the rise of deepfakes, the incorrect use of facial recognition and over-personalization. With a human-in-the-loop system, organizations can identify the points at which such AI technologies could take an unethical turn and steer them back onto the right path.

It’s clear that 2020 holds huge promise for the use of AI within an enterprise. However, businesses and consumers must understand that the fast-emerging AI capabilities aren’t a solve-all black box that can be blindly trusted. Correcting these misconceptions will be crucial for companies as they continue to leverage the power of the technology. With systems designed for continuous human input, enterprises can use AI to resolve roadblocks and deliver value in 2020 and beyond.