Adobe’s Generative AI Tools Face Criticism, Prompt User Concerns

Adobe’s digital imaging software has once again made headlines, this time among photographers who are expressing concern over the performance of the company’s generative AI tools. Initially unveiled in May 2024, the Firefly-powered Generative Remove tool was praised for its capabilities, operating effectively in many scenarios. However, ever since the official rollout of the tool moved beyond beta, its functionality seems to have taken a questionable turn.

Following recent updates that aired at Adobe MAX earlier this month, many users of Adobe Lightroom, Camera Raw, and Photoshop reported a drop in the quality of the generative AI features that they heavily rely on. This sentiment was echoed by various users on forums and platforms, with one user expressing despair over the lack of quality control, stating that the textures rendered previously have been replaced with results that appear smudged or artistic in nature. Some instances featured bizarre artifacts, such as animals appearing unexpectedly in pictures or additional limbs showing up where they shouldn’t.

PetaPixel’s own testing has aligned with these user experiences, highlighting a troubling pattern where Adobe’s generative AI technologies, designed to streamline photo editing, are instead hindering the workflow of many photographers. The generative tools, including Generative Fill and Generative Expand, aimed at enhancing editing speed and precision, have arrived laden with inconsistencies. A professional photographer, who uses generative fill thousands of times each day, lamented that despite previous versions achieving a near-perfect success rate, the updated tool now yields bizarre additions to images, such as unexpected rabbits or letters appearing on people’s faces. Their success rate, they shared, has plummeted to a disheartening 5-10% from the prior 90-95% range.

The photographer’s frustrations resonate with many in the community, who are experiencing similar mismatches in image textures and strange unintended placements of objects. The Generative Remove tool operates under the premise that, when removing pixels from an image, new pixels must seamlessly blend in with the background. In its current state, users report that this transition has become erroneous, leading to distractions being replaced with equally noticeable alterations.

While generative AI tools can drastically reduce the hours spent on labor-intensive tasks, the present experiences of photographers contrastingly illustrate a slowdown, with some users resorting back to traditional editing methods to manage their workflow. Adobe employee Terry White weighed in on the challenges of using Generative Remove, advising users to meticulously brush over the entirety of the object they wish to erase, including shadows and reflections. While this may improve results, the fundamental issues of the tool remain, leaving many users feeling conflicted about the evolving technology.

Despite the discontent, White acknowledged that improvements are still on the horizon. Adobe is actively listening to user feedback and making adjustments, an indication that they recognize the growing pains of integrating advanced AI into creative workflows. However, this inconsistency in results exposes a broader concern: the infancy of generative AI technology itself, which does not follow the linear path of improvement users may expect. This unpredictability poses significant challenges for photographers who place their trust in AI tools to enhance their productivity and creative output.

The situation has sparked conversations around the effectiveness of AI tools in photography; it is clear that while tools like Generative Remove and Fill have the potential to deliver outstanding results, they are not without flaws. As such, the quest for perfection in artificial intelligence-driven editing continues, with many practitioners still relying on traditional methods and acknowledging the limitations of current technology. For now, the journey remains a balance of AI innovation and the hands-on expertise that has characterized the field of photography for generations.