Loading article…
Google ImageFX debuts with Imagen 2, offering free text‑to‑image generation, expressive chips and safety watermarks in the US, Kenya, NZ and Australia.
Google unveiled ImageFX, a new text‑to‑image generator built on the Imagen 2 model, and made it available today in the US, Kenya, New Zealand and Australia through the AI Test Kitchen [1]. The rollout adds “expressive chips” that let users tweak lighting, mood or texture with a single click, while SynthID watermarks and IPTC metadata aim to flag AI‑generated images for downstream services.
| At a glance | |
|---|---|
| Tool | ImageFX (text‑to‑image) |
| Model | Imagen 2 (Google DeepMind) |
| Availability | US, Kenya, New Zealand, Australia (English only) |
| Launch date | 1 Feb 2024 (public rollout) |
ImageFX’s interface centers on “expressive chips,” a UI element that lets creators explore adjacent dimensions of a prompt—such as lighting or texture—without re‑typing the entire description [1][4]. Google says the chips accelerate creative iteration, a claim backed by internal testing that highlighted the importance of rapid prompt variation for user discovery [4]. All outputs carry a SynthID digital watermark and IPTC metadata, which Google asserts will help search and browser tools identify AI‑generated content when it appears online [1].
Imagen 2 also powers new generative features across Bard, Search, Ads, Duet AI and Vertex AI, positioning ImageFX as the visual front of a broader push to embed high‑quality AI imagery throughout Google’s products [1]. The company notes that Imagen 2 reduces artifacts and improves photorealism compared with earlier models, a technical upgrade that underpins the safety guardrails designed to block violent, offensive or sexually explicit outputs and to limit depictions of named individuals [1]. Google reports “significant investments” in training‑data safety and extensive adversarial testing to enforce these filters [1].
ImageFX joins a crowded market that includes Midjourney, DALL‑E 3 and Adobe Firefly. Unlike many rivals that charge per image or require a subscription, Google offers ImageFX free of charge and integrates it directly into its experimental Labs ecosystem, promising regular updates without additional cost [1]. The addition of expressive chips differentiates the tool by emphasizing rapid, UI‑driven exploration rather than raw generation speed alone.
ImageFX marks Google’s latest effort to democratize high‑quality generative imagery while embedding safeguards directly into the user experience, setting a benchmark for how large AI firms balance capability with responsible deployment. The open question remains how quickly Google can scale the tool globally without compromising the safety standards it touts today.
Coverage is mostly measured — 149 of 160 reports stay neutral.
Every Monday — the token unlocks, Fed dates & catalysts set to move crypto and markets this week. So you’re never blindsided.
Free · 3-min read · one-click unsubscribe
AI-assisted synthesis by the TrendWatcher Editorial Desk · sourced from 4 outlets · Jun 25, 2026 · How we report
AI Mode is a chatbot interface that uses Gemini 3 to generate answers to queries, replacing the list of links with a summarized response that still provides links for further reading.
AI Mode began rolling out to users in the United States in May 2025 and appears as a button on the Google Search page and app.
Critics say AI Mode could dramatically reduce click‑through traffic to websites, threatening the revenue model that many publishers rely on from organic search.
Google states that its AI tools still send billions of clicks to sites, that they increase the diversity and quality of traffic, and that the web is thriving with a 45% growth in content over the past two years.
Analyses have found that AI Overviews reduce click‑through rates by 30% to 70% and that about 60% of searches now end without a user clicking a link.