The Future of Emoji Communication
Why this needs a grounded rewrite
A lot of "future of emoji" content covering the broader kaomoji and emoji landscape leans on speculative technology that either never shipped or has nothing to do with how people actually communicate — holographic emoji, neural-interface emoji control, blockchain emoji ownership. None of that reflects what's actually happening in emoji and text-communication technology right now. What follows sticks to real, shipped features and documented roadmap items, with a clear line drawn between "this exists" and "this is a reasonable extrapolation."
What has actually shipped: AI-generated custom emoji
The most significant real development in emoji technology recently is generative, on-device custom emoji — not a hypothetical, but a shipped feature. Apple announced Genmoji at WWDC in June 2024 as part of its "Apple Intelligence" suite, and launched it broadly with iOS 18.2 in December 2024. Genmoji lets a user type a text description (or combine it with an existing emoji, or a photo of a specific person) and generates a custom, emoji-styled image on-device, which can then be used inline in text, as a sticker, or as a message tapback/reaction. This is a genuine category shift: for the first time, "emoji" isn't limited to a fixed, standardized character set voted on by a committee — it can be generated per-message, per-person, on demand.
This matters for how people communicate day to day: it moves highly specific, personal visual jokes (previously the domain of custom Discord/Slack emoji uploaded by community admins) into mainstream personal messaging, without needing image-editing skill or a shared server culture to access them.
What has actually shipped: the standardized Unicode roster keeps growing, on a predictable cadence
Separate from AI-generated emoji, the official Unicode-standardized emoji set continues to expand on roughly an annual cycle, through the same formal proposal process that added the 🤷 shrug and 🤦 facepalm emoji back in 2016. Unicode 17.0, released in September 2025, added a further batch of new emoji and sequences, with platform rollout continuing into 2026 as Apple, Google, Samsung and others build support into their own font releases (each platform ships its own visual design for every standardized emoji, on its own schedule, which is why a brand-new Unicode emoji sometimes appears as a placeholder box on one platform before another). This dual-track system — a slow, formal, cross-platform standardization process alongside fast-moving, platform-specific AI generation — is likely to be the lasting shape of "emoji" going forward, rather than one replacing the other.
How a new standardized emoji actually gets approved, step by step
It's worth understanding the formal process behind the "annual cycle" mentioned above, because it explains why standardized emoji move so much slower than AI-generated ones. New emoji are approved by the Unicode Consortium, a non-profit organization whose voting members include major technology companies — Apple, Google, Microsoft, Meta and others among them — through a dedicated Emoji Subcommittee. Anyone can submit a formal proposal, but Unicode's own published guidance sets a genuinely high bar, evaluating candidates against a documented set of selection factors: whether the concept is already representable with an existing emoji or combination (a strong reason for rejection), how distinctive the proposed image would be at a small size, whether it completes an existing category with an obvious gap, and evidence of likely real-world usage frequency, often supported by search-trend or social-media usage data submitted with the proposal. Proposals are also explicitly evaluated against a list of disqualifying factors — overly specific concepts, faddish or trend-dependent references likely to date quickly, and exact logos or branded imagery are all called out as reasons a proposal gets turned down. This is precisely the multi-year bar that delayed 🤦 and 🤷 until 2016 despite obvious everyday demand, and it's the same bar every future standardized emoji candidate has to clear, which is why the standardized track moves in years, not weeks.
The Unicode version history behind that "annual cycle" claim
The cadence isn't a loose approximation — it's visible directly in the Unicode Consortium's own version records, which tag every emoji with the exact release that introduced it. The original 2010 smiley block (Unicode 6.0, internally tagged emoji version E0.6) shipped the first wave of face emoji, including 😀, 😂 and 😒. E1.0, tied to Unicode 8.0 in 2015, added 🙄 (Face with Rolling Eyes). E3.0, part of Unicode 9.0 in June 2016, is the release that finally added 🤷 and 🤦 alongside Selfie and Fingers Crossed — with gendered (zero-width joiner) variants like 🤷♀️ following a few months later at E4.0 the same year. More recently, E14.0 (Unicode 14.0, September 2021) added 🫤 Face with Diagonal Mouth, and E17.0 is the current cycle referenced above. Reading the roster this way makes the "roughly annual" claim concrete rather than approximate: new emoji versions ship most years, but a genuinely new everyday reaction emoji — something filling a gap as obvious as 🤷 or 🤦 did — is much rarer than the yearly cadence suggests, since most releases add niche objects, flags or skin-tone/gender sequence combinations rather than new core emotional expressions.
Emoji Kitchen: the other precedent for combinable, generated emoji
Genmoji wasn't the first mainstream attempt at expanding emoji expressiveness beyond the fixed standardized set. Google's Emoji Kitchen, built into the Gboard keyboard and launched in 2020, lets users combine two existing standardized emoji into a single blended sticker — pairing, say, a fire emoji with a heart emoji to produce a unique combined image. Unlike Genmoji, Emoji Kitchen isn't generative in the AI-image-synthesis sense; it draws from a large pre-built library of artist-designed combinations rather than creating a wholly new image from a text description. It's a useful middle data point in the trajectory from "fixed standardized set" to "on-demand AI generation": Emoji Kitchen already broke the assumption that emoji has to be a single, fixed character per concept, years before Genmoji broke the assumption that emoji has to be pre-designed at all.
Custom emoji in workplace tools predate all of this
It's also worth noting that platform-specific, non-standardized emoji are not a new idea introduced by AI — Slack, Discord and similar workplace and community chat tools have allowed users to upload their own custom emoji, visible only within that specific workspace or server, since early in each platform's history. A team's inside joke turned into a custom :this-is-fine: emoji, usable only by people in that Slack workspace, is functionally the direct ancestor of today's AI-generated, hyper-personal Genmoji — the difference is entirely in how the image gets created (manually designed and uploaded by a person, versus generated on the spot from a text prompt), not in the underlying concept of "emoji specific to my group that no standardization committee will ever formally recognize." Seen this way, AI-generated emoji is less a break from precedent and more an automation of something communities were already doing by hand for years.
The accessibility layer nobody's future-of-emoji coverage mentions
Every standardized Unicode emoji has an official short name, defined through Unicode's Common Locale Data Repository (CLDR), which screen readers use to announce the character consistently — 🤷 is announced as "person shrugging" regardless of which vendor's visual design is displayed, which is precisely why the meaning stays stable even though the artwork varies. This system depends entirely on emoji being drawn from a fixed, labeled set. AI-generated custom emoji have no equivalent mechanism: a Genmoji generated from an arbitrary text prompt has no standardized accessible name a screen reader can announce consistently, because by definition no two generated images are guaranteed to be the same or to have been reviewed for an appropriate label. As generative emoji become more common, this is a genuinely unresolved gap between how accessible standardized emoji are today and how accessible AI-generated ones currently aren't — worth flagging plainly rather than glossing over in coverage focused only on the exciting parts of the new feature.
What's a reasonable extrapolation, clearly labeled as such
Based on the trajectory of what's already shipped, a few developments seem like reasonable near-term extensions rather than speculation:
- More platforms building their own generative-emoji features. Given Apple's Genmoji launch, competing platforms building comparable on-device or cloud-based custom emoji/sticker generation is a plausible, low-risk prediction rather than a stretch.
- Continued gradual standardization of representation. Unicode's emoji process has consistently, year over year since roughly 2015 (when skin-tone modifiers were added in Unicode 8.0), expanded representational options — gender-neutral figure variants, disability-representing emoji (introduced starting Unicode 12.0 in 2019, including a wheelchair symbol, guide dog, and hearing aid), and similar additions. Continued incremental expansion in this direction is consistent with the pattern, not a leap.
- AI-assisted emoji suggestion becoming more contextual. Predictive emoji suggestion based on message content already exists on most major keyboards (Gboard, iOS keyboard); deeper integration with on-device language models for more accurate context-matching is a natural extension of existing, shipped functionality rather than a new category.
A timeline of notable Unicode emoji milestones
| Release | Year | What it added |
|---|---|---|
| Unicode 6.0 | 2010 | The original standardized emoji set, drawn from existing Japanese carrier pictographs |
| Unicode 8.0 | 2015 | Fitzpatrick-scale skin tone modifiers |
| Unicode 9.0 | 2016 | 🤦 facepalm and 🤷 shrug, following a public proposal campaign |
| Unicode 12.0 | 2019 | Disability-representing emoji, including a wheelchair symbol, guide dog and hearing aid |
| Unicode 15.0 | 2022 | Continued expansion of the general roster |
| Unicode 16.0 | 2024 | Further roster additions, alongside Apple's Genmoji launching the same year as a separate, non-standardized track |
| Unicode 17.0 | 2025 | The most recent standardized release, with vendor rollout continuing into 2026 |
The pattern across every row: the standardized track has added new emoji roughly annually since 2010, each release going through the same formal proposal process described above, with vendor-specific artwork and platform rollout trailing the actual Unicode approval by months.
AI-generated content and the growing value of provably human text
There's a less obvious angle worth raising as AI-generated content — written text, images, and now custom emoji — becomes more common across the internet: symbols that are cheap and mechanical to produce at scale don't automatically become less useful, but the specific things people reach for as markers of a genuinely human, deliberate touch can shift. A hand-placed ¯\_(ツ)_/¯ in a message costs the sender nothing to type (once copied) but still requires an actual person choosing, in that specific moment, to reach for a very old, recognizable piece of internet shorthand rather than generating something new. Whether that distinction ends up mattering to readers as AI-generated custom emoji become more common is genuinely speculative and shouldn't be overstated as a prediction — but it's a reasonable enough dynamic to flag as one more advantage sitting alongside the more solidly established one from earlier: the text shrug's zero dependency on any rendering engine, any AI model, or any platform-specific feature at all.
Where text kaomoji fit into this AI-generated future
It's worth being specific about why text-based kaomoji like ¯\_(ツ)_/¯ are structurally different from everything above, and unlikely to be displaced by any of it. Genmoji and AI-generated stickers are, technically, images — they require image rendering support, don't work in plain-text contexts like terminals or code comments, and (as of Genmoji's current implementation) are tied to specific hardware capable of running on-device generative models. The text shrug requires none of that: it's plain Unicode characters that remain copyable in a 2004 web forum, a 2026 phone, a code comment, or a rich-text message. Its glyph shape and spacing can still vary by font, and Markdown can consume its backslash unless it is escaped. Rather than AI-generated emoji making kaomoji obsolete, the two solve different problems — expressive, personalized visual communication on one hand, and dependency-free plain-text communication on the other. See the full text shrug vs emoji shrug comparison for exactly where each one wins.
Who actually tracks and reports on all of this
Emojipedia, founded by Jeremy Burge in 2013 — the same person credited with co-authoring the 2015 facepalm and shrug proposal referenced earlier in this piece — has become the de facto reference site the tech press turns to for tracking new emoji releases, vendor-specific design changes, and platform rollout timing, and it's now owned by Zedge after an acquisition. Burge also created World Emoji Day, observed every July 17th (the date chosen because it's the date shown on the calendar emoji 📅 itself), which has since become an annual moment for platforms to announce new emoji features and for outlets to publish usage-trend coverage. Emojipedia's own tracking is generally considered the most reliable independent source for "what's actually shipped where" specifically because platform vendors don't always publish their own detailed emoji-support changelogs, making a dedicated, consistently updated third-party tracker genuinely useful rather than redundant.
What's genuinely uncertain, and worth being honest about
A few open questions don't have a clear answer yet, and any confident claim about them should be treated skeptically:
- Whether AI-generated emoji become interoperable across platforms. Currently, a Genmoji generated on an iPhone is essentially an Apple-ecosystem image asset, not a cross-platform standardized character the way a Unicode emoji is. Whether generative emoji ever gain the kind of universal cross-platform guarantee that standardized Unicode emoji have is genuinely unresolved.
- Long-term accessibility implications. Standardized emoji have defined accessible names that screen readers announce consistently (see the accessibility comparison for how this already differs between the text shrug and the 🤷 emoji). It's not yet clear how AI-generated, non-standardized custom emoji will be made consistently accessible, since each one is unique rather than drawn from a fixed, labeled set.
- Whether the formal Unicode proposal process itself changes pace. The annual review cadence has held steady since 2010; whether growing pressure from faster-moving AI-generated alternatives eventually pushes the standards body to move faster, stay deliberately slow as a differentiator, or stay roughly as-is isn't something the public roadmap currently answers either way.
- Whether generative emoji end up governed at all. Standardized emoji go through public proposal review specifically to filter out offensive, confusing, or poorly-designed additions before they ship to billions of devices. On-device AI generation has no equivalent review step by definition, since each output is unique and generated in the moment — what content-moderation guardrails, if any, become standard practice across vendors is an open question rather than a settled one.
None of this uncertainty touches the durability question covered separately in why the shrug emoji went viral and stayed that way — that piece's argument holds regardless of how the AI-generated side of emoji technology develops.
The steady middle: what probably won't change
Despite the real advances in AI-generated and expanding standardized emoji, the core reasons text kaomoji have persisted for two decades — broad font support, zero dependency on image support, and workability in plain-text environments like code and terminals — aren't affected by any of the developments above. If anything, as visual communication options multiply and fragment across platform-specific generative tools, the text shrug's cross-platform reliability becomes more distinctive, not less. The Unicode Consortium's official emoji version table documents the release history referenced above. For the full history of how the shrug got here, see history of the shrug emoji.