Silhouette of a woman sitting in a grassy field with arms raised, an AI network in the air around her as the sun rises

AI is great at speeding up processes, generating deeper insights, reducing costs, and increasing sales. So say dozens of interviewees from twenty countries and multiple companies, according to our ongoing research. But this isn’t actually news—areas and industries all around the world are already leveraging AI with great results.

On a grander scale, localized implementations in manufacturing, healthcare, and infrastructure could have sweeping positive impact. Instead of building new roads to reduce traffic, for example, we could use AI to coordinate private and public transport routes on existing roads, saving time and construction costs.

But before we can move ahead with change, we have to get people onboard. Data suggest that to create an AI-positive mindset in the greater population, investing shouldn’t just target R&D—it should target learning, too. China, the country leading the charge towards an AI-friendly future, has set a goal for 100% of school children to become competent in computer science by 2025.

COVID-19 has increased the need for us all to internalize the value of AI, and luckily, global institutions are catching on. The worldwide need for recovery has accelerated investments in AI everywhere. But it won’t be a quick switch.

Here are three things that all of us—investors, institutions, and laymen alike—need to understand about AI to leverage its evolution.

Tech Takes Time

Though technology has a reputation for speed and acceleration, technological breakthroughs take decades to make a difference to the aggregate economy. Cars started appearing at the turn of the century, but it would be years before complementary investments such as asphalt roads, traffic lights, and mechanics gave value to the motor industry.

What does this equate to when we’re talking about AI in 2020?

In the same way that cars benefitted from roads and mechanics, AI will need infrastructure and data scientists. But that’s not all.

Consider that AI is fed by data in the same way a car is fed by refined oil. That means we need data, zillions of data from multiple sources, all refined in the same format. Complementarity creates coordination of the ecosystem. The data generated by economic indicators, patent registration, career consultants, and even innovation courses at business schools must be collected, analyzed, and shared.

The more people who invest in AI, the quicker AI will evolve, and the sooner we can put it to use out there in the world.

Variety Is Everything

A coordinated ecosystem for AI depends on varied data, but variety in data sources is vastly helpful, as well. That is to say, we need different people investing in AI in different ways.

We need both private and public investments—as per cars, we need private gas stations as much as public roads. For AI, that equates to private organizational capital for platforms or sensors that collect data, as well as public physical capital from institutions to define cooperation standards and promote insights.

Without variety, we’ll end up with biased data, and that would do more harm than good in the long run.

Distribute Far and Fairly

Even with time and varied data, a fair distribution of resources and capabilities to develop AI (and the eventual product of that development) is absolutely necessary. For that, we need coordination.

Any transition that involves millions of people has to be coordinated—from investment to implementation—between private entities and governments. Overlooking or trivializing this aspect of AI implementation could have serious repercussions. This is already being proven by the AI Economist, a Salesforce research project unveiled in April 2020 that applies machine learning to economic models. Researchers are well aware of how marginalized groups will be difficult for hyper-rational AI to understand due to the simple fact that those groups are in the minority. There is even a concern that AI may sacrifice them for what it may perceive to be the greater good.

In short, machines cannot interpret data or distribute AI benefits fairly on their own. For AI to best serve everyone, everyone must be involved from early stage development to aggregate implementation.

Bottom Line: Put Empathy to Work

As much as we know that AI can make a huge impact on society, we also know that humans are required to develop and maintain it. AI doesn’t have empathy, and it can’t interpret data.

Fortunately, people are very good at that.

If we have patience, ensure varied data, and coordinate to assure the fair distribution of resources, the world will be primed to reap the benefits of AI throughout the pandemic and beyond. While it may seem like a lot of work lies ahead, we already have working models for technological innovation. The three-pronged technology-strategy-psychology model, for one, will help ensure we keep a balanced view as AI develops. We also have a model for implementation based on three axes: industry-spanning research, complex system thinking, and powerful decision making.

With all the negative changes COVID-19 has brought, data show that motivation is high for transformation. Now is the moment to leverage AI through global coordination and education.