Can AI Wearables Replace Phones?

In January, The Information reported that Apple is speeding to make a wearable AI pin the size of an AirTag. By February, Apple’s yet-to-be-made AI wearables expanded to include glasses, pendants and AirPods fitted with cameras. Given Apple’s reputation for being the most sluggish big tech company in offering AI services to its users, the timing of the announcement made it a declaration of competition with OpenAI’s wearable. For context, Sam Altman and Jony Ive announced their first physical AI device in May 2025, with launch set for the second half of 2026. 

The AI wearable rivalry has since expanded to include Meta and Google. Apple will be competing with Meta on smart glasses (where Meta has first-mover advantage after sinking billions). Notwithstanding, big tech players are not the only ones scrambling for the $62B market opportunities.

The core principle of AI wearables is to augment human capabilities and free up time for other personal aspirations. Too much screentime, especially on phones, has generally reduced our presence of mind, health and quality of life. The plan is to reduce digital interactions via screens and reroute them through voice commands, audio feedback, and gestures. Newer wearables are already going screenless. Can the same companies that got us addicted to our phone screens break the addiction? What’s in it for them?

History

Difference between smart and AI wearable devices 

Wearable smart devices are electronic devices worn on the body that collect and process data and communicate wirelessly — either by connecting to the internet or by pairing with another device. They are a subcategory of wearable technology.

There’s a slowly disappearing thin line of difference between what a smart wearable device is and what an AI wearable device is. Many smart devices already have AI capabilities to some extent. The distinction only became necessary because AI costs have been sunk to humanise the data originally collected by these devices. Essentially, wearable devices came first, collected a ton of data and generated some insights without the capacity to learn from the data, adapt, and act autonomously, which AI wearables now have the capacity for. 

Wearable AI Market Size, Share & Industry Analysis

Fortune Business Insights reports that the AI wearables market size is projected to grow from $61.51 billion in 2026 to $359.32 billion by 2034. With AI disrupting almost every industry, research has focused on developing AI devices for everyday use. Since AI technology is still very nascent, there still remains plenty of room for innovation.

Based on product type, the market is currently segmented into smartwatches and fitness bands, smart eyewear, smart earwear, smart clothing, and other categories (smart rings, smart patches).

By application, it is segmented into consumer electronics, healthcare, automotive, military and defence, media and entertainment, and other categories.

Market opportunities

The race to pioneer the production of any segment of AI wearables obviously stems from the market size and profits, and it also comes with the promise of being the next iPhone of the era. Wearable AI technology is still in its early days of development, with every part of the human body as the next frontier. A glimpse of demand for these products can be seen during the pandemic – consumer demand for health monitoring devices showed steady growth, in spite of supply chain disruptions.

The Place of Screens

Before the introduction of AI technology, smart devices monitored health and wellbeing by tracking activity levels and displaying the data on screens. This fuelled the adoption of healthy lifestyles. Then, AI tech (ML algorithms) helped take smart wearable devices beyond just monitoring. It provided predictive analytics to identify potential health issues, suggested intelligent remedies, and used monitoring data to proffer individualised features.

Though the main selling point of AI wearables is personalised solutions, these solutions were often displayed on screens. Now, newer devices are not just planning to do away with screens but are already doing so. The marketing for Google’s new Fitbit Air includes phrases like, “… screenless design built so you can live in the moment”, and “… stay notification-free”. Google is able to deliver on these promises because insights from your health data are sent to your phone via the Google Health app.

The argument for going screenless in the traditional sense — AI wearables eventually need a visual medium to display information — is that phones were not meant for the more holistic, simultaneous interactions AI agents are capable of. This makes phones very restrictive. In order to reach full service potential, new AI-native devices need to be everything that a phone is, without the bulkiness, not distract the user from being “in the moment”, and still be capable of more. An example is AI glasses that can display information about your environment on one lens while the other is free, equipped with a microphone for voicing commands instead of whipping out a phone to type, and can take pictures with gestures or voice commands. It also needs to be intelligent enough to take notes, record audio of interactions, and, like an experienced assistant, accurately anticipate your needs based on what it knows about you. 

Even with the promise of allowing us to be more present, current AI wearables on the market are either phone wannabes or glorified trackers that still need a phone’s screen to function optimally.

AI Wearables That Came Before – Failures

On the 21st of May, 2025, Sam Altman and Jony Ive announced io – a company born from OpenAI’s need to make hardware devices. Thus, introducing the AI hardware frenzy. Before the announcement, some startups tried to incorporate AI into screenless smart devices. Their products failed to satisfy user expectations and did not generate the same amount of fervour big tech companies are displaying towards AI wearables now. From the products that have come before, to the ones in the making, one promise rings true: phones screens are going to be made obsolete.

We cannot discuss the failures of earlier AI wearables without touching base with current requirements for AI wearables.

Requirements for today’s AI wearables

Input and output modes:

Holographic visual display (because sight is an important sense organ that embodies great depth in human interactions), audio for voice prompt and feedback, and gesture as command. Cameras are inevitable for visual input. At the end of the day, we are all going to be cameras. Glasses are the perfect wearable AI devices because they sit near your eyes, ears and mouth. 

Internet connectivity:

Although much of the agentic AI intelligence and reasoning of wearables will be taken care of by mobile edge computing advancements, their agentic abilities still require always-on streaming. 5G is no longer enough. 6G will power the ultra-low latency required for real-time streaming to the cloud and instant feedback.

Security: 

Some of Google’s Android Show I/O Edition 2026 updates include: “personal browsing assistant”, turning a webpage into an informative infographic, auto browsing and auto-updating your online orders with lifestyle changes/needs”. These capabilities leave plenty of room for AI security challenges like prompt injection. Cybersecurity protocols have to be extra tight since these agentic sessions are completed via the internet.

Battery life:

Part of AI wearables’ philosophy is discreetness — something that is a natural extension of you. This shows in the forms they have taken: pins, glasses, earwear, and wristwear.  Their size means batteries for wearables have to be miniaturised, lightweight, flexible, and last all day, at the very least.

These factors directly and indirectly contribute to the seamless user experience expected of a non-disruptive yet highly needed device.

Case Study of Failed AI Wearable Brands

Humane Inc., founded by two former Apple employees, launched the AI pin in November 2023. Its first batch of products shipped in April, 2024. In February 2025, the company sent a mail saying they were winding down. How did their promise of “screenless so you stay in the moment” lead to a one-way ticket to startup graveyard?

The magnetically attached pin was meant to replace the phone, but it failed in places where the phone had maximised potential. Cost wise, its  $700 initial purchase expense and an extra $24 monthly subscription didn’t work. Most smart devices marketed with the slogan “AI-enabled” come with the added burden of monthly subscriptions: to access the AI insights from the raw data the device collects and to cover data center costs. Outside of power users, will normal users adjust to this service model?

User experience was poor – gesture controls were frustrating and not familiar. Unless you’re offering exclusivity or luxury, users prefer intuitive products.

Battery performance was also poor. It would have been overlooked because of its small size, but it also overheated.

What Has Worked

Mobile edge computing is aiding newer products to bypass the cloud via on-device processing. The result is faster processing time, multi-layered sensors, continuous connectivity, extra security, and longer battery life.

Policy and Privacy Opportunities/Concerns

Given that every tool, regardless of the intent for making them, can be abused or exploited, what measures are already in place or should be put in place to safeguard wearable tech consumers?

News of a Russian blogger who used smart glasses to secretly film consensual sexual acts with women across Africa is still unfolding. Another example of a product flopping once users became aware of its 24/7 surveillance capability is Ring’s Super Bowl commercial.  The smart home security company announced its new AI-powered neighbourhood surveillance feature in the guise of helping find lost pets. Users were not having it.

These privacy concerns cannot be hushed away since anyone can be recording at any time. How prepared are we to navigate a world where our bodies are tech hubs? A world where casual conversations with friends are recorded events. We need rules to ensure we get the benefits of wearable AI without losing trust. The de facto rule is to assume that you are being recorded, but the government can do more.

Wearables complicate matters because they process data on-device, continuously collect sensor data, and make context-aware decisions. Rules need to cover not just how data is stored but also how it’s collected and how decisions are made in real time. Governance policies must include expiration dates for data permissions so data doesn’t stay around forever. 

AI Wearables and the Future of Human Experiences

Wearable technology is already changing the way we learn, work, socialise and even create memories. Embedded in these devices are ‘upgraded’ AI assistants that are supposedly smarter than the ones on our phones. They are designed to become even more personalised for a better user experience. Current and projected capabilities of AI beyond being a basic assistant include anticipating the user’s needs. Let’s not even get into the security issues surrounding such a promise and how much of our privacy and willpower we would be signing over with this bargain.

After glasses, implants are the obvious next step for wearables — computers have a trend of getting smaller with each iteration. If phones, as sleek as they have become, and can be put away to focus, are considered inconvenient and addictive, I can imagine having my addiction be something I have to wear as a second set of eyes or embedded in my body. Implants seem more convenient. They are also perfect for when thought-as-input mode (where the AI reads your brain waves instead of voice prompts) becomes more ubiquitous.

 “Unlike a laptop or phone, a wearable is designed to occupy the same physical space as a learner and to move with them through school, home and fieldwork. That means continuous contextual data — gaze, location, ambient audio, short video clips, and biometric readings — may become available for adaptive tutoring, classroom management, and formative assessment. The promise is personalised in situ feedback: a wearable that detects a student’s confusion during a lab and provides a hint, or that captures a teacher’s explanation and produces a searchable summary. But the presence of persistent sensing also creates new structural risks. Surveillance-style deployments can chill discussion, undermine equitable access, and transform informal peer interactions into recorded artifacts. The policy imperative is to design norms and procurement rules that preserve pedagogical gains while preventing data capture that corrodes trust.” – The Economy Review

Going screenless might just be a marketing hype, or not. At the end of the day, a visual display medium is needed for digital interactions.