My mission for more than two decades has been to help artificial intelligence (AI) work for the masses, and I truly believe in AI’s potential to make our lives healthier, happier and more productive.

Of course, AI comes with certain challenges (like any emerging technology), especially as companies more fully operationalize AI. In fact, there are at least three significant AI trends already on the move this year — and I’m closely watching how these shifts will help businesses continue to navigate AI hurdles like removing bias and building trust.

Trend 1: Actioning AI ethics and governance 

For years, there’s been discussion about eliminating bias from AI models. These concerns are not only prominent among industry professionals — they’re increasingly arising in mainstream media outlets as well. For example, NBC recently aired an episode of “American Auto” that centered on a self-driving car’s failure to recognize and brake for people of color.

In 2022, we’re seeing conversations about AI ethics and bias mitigation transition from abstract frameworks into real-world practices. This evolution is largely powered by emerging startups that  provide AI monitoring and governance solutions for businesses. Now, a big question mark for AI-driven companies is whether to outsource machine learning performance monitoring to companies like Credo, Fiddler and Arize AI, or build out internal capabilities to validate, monitor and analyze machine learning models.

My advice: Don’t overthink this decision. At Smart Eye, we have largely implemented checks and balances in house, but I look forward to exploring potential partnerships with emerging companies in the governance space. However, if your organization currently lacks the in-house expertise to properly operationalize AI ethics and governance, go ahead and bring on a partner who can. Many third-party solutions can implement systems that can train, validate and analyze the efficacy of your AI systems, and at levels that are difficult to achieve with data offered by your customer base alone. Start simple as well, perhaps by monitoring the diversity of your training and test data, or biases present in your inference results. Then you can work off of this information to intervene as required.

Over time, you can increase efforts and adopt more tools and capabilities to help eliminate bias and add model explainability. From my perspective, the important thing is that your organization is bringing AI ethics and governance into action now.

Trend 2: Increasing AI’s role in hybrid workplaces

According to recent research from Microsoft, more than 70% of workers globally want flexible remote work options to continue. I know that’s how I personally feel. Hybrid work environments are here to stay — even at my company, we’re figuring out what hybrid looks like in practice, and how to get ahead of employee burnout. But one thing I’m sure of is that AI will continue to drive innovation in the future of work. 

We’ve seen a recent rise in organizations embracing collaboration and workspace tools designed to boost engagement and happiness levels. As a next step, layering in AI can help you learn so much more about how your team is doing. For example, I’m a big fan of startups like Read AI, which monitor meetings in real time to gauge who speaks the most, the overall sentiment in the “room” and other nuanced behaviors. Over time, gathering this information helps leaders improve future experiences for employees, surface helpful coaching insights and uncover team members’ skills. 

The pandemic has accelerated the adoption of virtual and/or hybrid settings, and it’s great to see more tools coming to market that can quantify and support social and emotional intelligence in organizations. Upgrading how you engage with employees has powerful mental health implications, too. Rightfully so, building out more helpful mental health resources remains a top priority for organizations in 2022 — nearly 40% of employers expanded mental health benefits during the pandemic. I’m eager for more large-scale deployment of AI systems that better quantify an individual’s mental health needs and can provide just-in-time support — stay tuned for more updates on that front!

Trend 3: Exploring AI and Web3

A third trend I’m keeping my eye on is the intersection of AI and the emerging world of Web3, crypto and NFTs (non-fungible tokens). 

One obvious area where AI is being applied is in synthetic data — otherwise known as artificially created data. In my world, we turn to synthetic data all of the time to power deep learning models and train generative models — without having to dedicate massive amounts of time and money to producing labeled diverse data sets. We’re slowly seeing the application of generative adversarial networks in Web3, where one can create thousands of unique, synthetic characters to populate the metaverse. This opens the door to engineering new user experiences, and even exploring new monetization and branding opportunities (think influencers in the metaverse!).

The same goes for NFTs. While NFTs themselves still feel quite new, there are a lot of opportunities to embed AI and make these digital assets more interactive. Imagine intelligent NFTs (iNFTs) that have natural language understanding, perceptual capabilities and computer vision, and therefore can engage audiences in conversations — an iNFT telling you about its “origin story,” for instance. This is definitely a space I’m watching.

No matter the trend, keep AI human

Amid divergent trends, there’s a consistent element in 2022: keeping humans central to the AI equation.

This goal is giving way to the emergence of a new technology category: human insight AI — AI technologies designed to understand, support and predict human behaviors within complex environments. This year, we’re already seeing human insight AI strategies leading to improvements, from the interiors of our cars to hybrid work environments … and maybe even in the metaverse. But no matter where human insight AI is applied, we will all be better off for it, and I can’t wait for these human elements to power all sorts of AI experiences in 2022 and beyond.