The work AI should really be doing, according to these pros (www.zdnet.com)

Fast Company asked nine leading designers what they want AI to actually do for creative work, and their answers converge on one idea: not an invisible autopilot, but a trustworthy creative partner. Designers want generative models (text, image, audio, video) to remove tedious, repetitive tasks—kerning, routine edits, rapid prototyping—while preserving the “friction” that sparks collaboration, mistake-driven discovery, and craft. Practical examples include people using chatbots like ChatGPT as idea partners and companies like Luma AI pushing “reasoning” video models (Ray3) and adaptive agents that can hold context and nuance for specific creators or teams. At the same time, they flagged wider ecosystem risks: a deluge of synthetic content, copyright and training-data disputes, and the tension between automation and meaningful human control. For the AI/ML community this translates into concrete technical and product priorities: build models that maintain context, support controllable creativity, and offer fine-grained human-in-the-loop controls and provenance. Developers should focus on reasoning-capable architectures, personalized agents that adapt to individual workflows, and interfaces that enable experimentation rather than top-down mandates. Success will be measured less by raw automation and more by ROI through adoption, trust, and preserved creative agency—three ingredients designers named as essential: experimentation, open-mindedness, and individuality.