3 A.I. Trends to Watch in 2022

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

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

For years, there’s been discussion about eliminating bias from A.I. 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 A.I. ethics and bias mitigation transition from abstract frameworks into real-world practices. This evolution is largely powered by emerging startups that provide A.I. monitoring and governance solutions for businesses. Now, a big question mark for A.I.-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 A.I. 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 A.I. 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 A.I. ethics and governance into action now.

According to recent research from Microsoft, more than 70 percent 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 A.I. 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 A.I. 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 monitors 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 percent of employers expanded mental health benefits during the pandemic. I’m eager for more large-scale deployment of A.I. 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!

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

One obvious area where A.I. 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 A.I. 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.

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