AI Has a Real VR Problem

Is Artificial Intelligence the new kid stumbling down the same rocky path as Virtual Reality?

When AI Met Reality (And Reality Won)

Artificial Intelligence burst onto the scene like a rockstar promising the world—and maybe your job—to its algorithms. From self-driving cars taking over highways to AI doctors diagnosing illnesses with a mere glance, the future looked automated and bright. But over the past year, the AI anthem hit a few sour notes.

While AI performs outstandingly in tasks like image recognition and language translation, it often falls over in real-world scenarios that don't fit neatly into its training data. As highlighted in The Economist in 2021, "What happened to the artificial-intelligence revolution?" underscoring AI's struggle with context beyond pre-defined parameters. Ethical concerns also loom large. MIT Technology Review sounded the alarm in "The problems AI has today go back centuries" (2021), pointing out biases that can seep into AI systems in marginalising communities. These examples and many more remind us that AI sometimes overpromises and under delivers—much like that treadmill collecting dust in the corner.

So, where do we go from here without pulling the plug? Some suggest narrowing our focus. In "The Unsexy Future of Generative AI Is Enterprise Apps" (Wired, 2024), experts argue for specialising AI applications—specifically in niche, B2B use cases—rather than chasing the elusive general-purpose AI. Improving transparency of AI reasoning and processes is also key. The Conversation article "A ‘black box’ AI system has been influencing criminal justice decisions for over two decades – it’s time to open it?up" (2023) advocates for making AI decisions understandable, highlighting that the lack of access to the data and other crucial information required for independent evaluation raises concerns around accountability and transparency of the decision process.

What's clear is addressing these concerns around confidence in capabilities and transparency around decision making is crucial step regaining public trust in AI—even if just to assure us these intelligent systems aren't plotting world domination.

VR's Wild Ride: From Hype to Horrible to Hope

Virtual Reality was supposed to be the next big thing—the technology that would make our wildest sci-fi dreams come true. We envisioned ourselves exploring fantastical worlds from our living rooms or having meetings as holograms (because who doesn't want to attend more meetings?). But the VR revolution hit some speed bumps.

High costs kept VR out of reach for many consumers during the disillusionment stage of the hype cycle. Even those who could afford it at the time found a lack of compelling content—a virtual world that's, well, virtually empty. Fortune pointed out both these challenges in "The fall and rise of VR" (2019) saying "a dearth of content" and "VR is typically too expensive" are the main barriers users were experiencing to justifying the content. Even today I personally have a Meta Quest 3 and all I use it for is Beat Saber not being compelled in justifying the price of many other dedicated VR games.

Of course, I can't mentioned challenges with the adoption of VR without mentioning the elephant in the room—user discomfort. Headsets are often clunky and many users experience motion sickness, especially with earlier models. Then again, it's difficult to make strapping screens to our faces a pleasant experience.

Yet, VR didn't fade into oblivion. Developers pivoted to create more affordable hardware, as discussed in "Headset technology is cheaper and better than ever" (The Economist, 2020). They also expanded VR's horizons beyond gaming, finding uses in niche industries and use cases (just like AI is being recommended to do now) Check out Oxford Medical Simulation for a great example of this. Enhancements in user comfort, display-movement synchronisation, and refresh rates of screens have helped create a more pleasant experience covered overall with promising results. In "Survey of Motion Sickness Mitigation Efforts in Virtual Reality" (IVRHA, 2024) Benton Lane discusses a range of improvements made in VR in recent years and how these improvements have lead to improved experiences and increased comfort using the technology overall. What's clear is VR is finding it's feet—which is literally a huge improvement.

Same Cycle, Different Tech

What does AI have in common with VR besides two-letter abbreviations and a knack for overpromising? Quite a bit, actually!

Both technologies skyrocketed to fame on waves of hype only to crash into the hard reality of user expectations. They grappled with accessibility—VR with its costly hardware, AI with its complex algorithms that few understand (or trust). Both faced a critical need to pivot toward specialised applications where they could genuinely shine.

This is fairly common with the adoption of new technology and is called the Gartner Hype Cycle, which typically follows five key phases.

1. Trigger – A technology breakthrough kicks of keen media and public interest.
2. Peak of Inflated Expectations – This media and public attention spins out of control building massive hype. Around this time and scores of failures and disappointments start to be discovered.
3. Trough of Disillusionment – Interest wanes as the new technology fails to deliver on everything that was promised.
4. Slope of Enlightenment – More instances of the technology's benefits start to crystallise and be better understood.
5. Plateau of Productivity – More mainstream adoption starts to take off.

Right now, AI is in the Trough of Disillusionment and entering the Slope of Enlightenment. VR is entering the Plateau of Productivity.

So, what's the lesson here? Managing expectations might just be the secret sauce for AI navigating the hype cycle successfully. VR adjusted by finding niche markets and improving user experience. AI can do the same by focusing on areas where it excels and being transparent about its limitations. After all, not every problem needs an AI solution—sometimes, a good old-fashioned algorithm will do just fine.

Can AI Climb Out of the Trough?

If VR can rebound from its stumble, there's hope for AI. By learning from VR's detour through the valley of disappointment, AI can chart a course that avoids the same pitfalls. Emphasising ethical considerations, enhancing transparency, and honing in on practical applications can help AI regain its footing.

Consider this an opportunity for growth rather than a setback. As tech leaders and innovators, embracing a realistic approach to AI implementation can pave the way for sustainable progress. And who knows? Maybe we'll finally get those AI-powered personal assistants that don't accidentally order 100 kilos of dog food when we ask for the weather forecast.

–

In the constant march of technological advancement, both AI and VR remind us that hype can be a double-edged sword. It's easy to get swept up in grand promises, but the real work happens when the spotlight dims.

By managing our expectations and focusing on down-to-earth, tangible, and user-centric solutions, AI can navigate through its current adoption challenges. So let's take a page from the VR playbook—not the one about making people dizzy—and steer AI toward a future where it not only impresses but also delivers.

After all, the journey ahead is a marathon. And sometimes, it's okay to walk—especially if it's through a virtual landscape without the motion sickness.