Photoshop Adds Markup Tool for AI Prompt-Based Editing

A New Way to Talk to Photoshop

Adobe is pushing Photoshop further into the era of generative AI with a new feature called Markup, which lets users draw directly on an image to communicate what they want changed, rather than relying solely on typed prompts 1. The tool is part of a broader rollout of AI-assisted editing capabilities and Lightroom-style controls that longtime Photoshop users have requested for years 1. Reporting indicates Adobe is not just adding a feature but fully leaning into prompt-based editing as a core workflow, launching a new AI-assisted editor that blends traditional pixel-level tools with generative instructions 3. The upshot is a more intuitive interface: instead of describing a desired edit in words alone, a user can sketch an arrow, circle a region, or scribble a rough shape to show the AI precisely where and how to intervene.

Part of a Bigger Industry Shift

Adobe's move lands amid a broader surge in image-generation technology across the tech industry. OpenAI's ChatGPT recently updated its image model to a version often referred to as Images 2.0, which produces pictures that are markedly more photorealistic and correspondingly harder for viewers to identify as AI-generated 4. Google, meanwhile, is expanding generative imagery into its core search product, adding AI image generation to AI Overviews and redesigning the Google Images homepage, with the rollout beginning over subsequent weeks 6. Independent monitoring confirmed that Google had begun actually surfacing AI-generated images within AI Overviews in Search, following an earlier announcement that the capability was coming 7. Together, these developments show generative image tools moving from standalone creative apps into everyday search and productivity surfaces used by millions.

Why Realism and Trust Are Becoming the Central Issue

As these tools grow more capable, the line between authentic and synthetic imagery keeps blurring, with real-world consequences. A recent example is Donald Trump's AI-generated image of himself horse riding alongside George Washington, which drew attention not for its technical quality but for an unintended symbolic misstep involving the horse's color, illustrating how AI-generated content can carry meanings creators may not intend or foresee 2. This kind of incident underscores a growing tension: as tools like Photoshop's Markup and ChatGPT's newer image model make generation faster and more realistic, the potential for confusion, misrepresentation, or symbolic missteps grows in tandem 42.

Creativity, Identity, and Everyday Use

Beyond politics and search engines, generative imagery is reshaping how ordinary users express themselves online. Social media has long been a space for experimenting with visual identity, and AI avatars, synthetic game worlds, and generative editing are becoming the newest layer of that experimentation, extending well past traditional selfies and edited photos 5. Adobe's Markup feature fits squarely into this trend: it lowers the technical barrier to sophisticated image editing, letting casual users direct powerful AI models with simple gestures rather than technical prompts 13.

The Bigger Picture

Taken together, these developments—from Adobe's gesture-based editing to Google's search integration and OpenAI's more realistic outputs—signal that generative image and multimodal AI tools are rapidly becoming embedded in mainstream software rather than remaining niche experiments. That shift raises the stakes for authenticity, misinformation, and creative norms, even as it opens new avenues for expression across professional design, search, and social media alike.