Ambition of Integrated AI Platform and the Cold Skepticism Brought by ''Automation of Creativity''
Adobe MAX 2025 conference keynote announced major expansions to the Firefly AI ecosystem. Key announcements: (1) Multi-model partnerships — Firefly can now use AI models from Google, Black Forest Lab, Topaz Labs, ElevenLabs (voice), and others beyond Adobe''s own model; acknowledges that specialized models outperform general-purpose models in specific domains; (2) Personalized AI models — users can upload their own images to create custom LoRA-style models generating images in their unique style; enables brand-consistent generation and personal style preservation; (3) AI Assistant in Creative Cloud — conversational interface integrated across Photoshop, Illustrator, Premiere; enables natural language instructions for complex multi-step tasks; (4) Video generation advances — Firefly Video expanded to generate up to 2-minute clips with improved temporal consistency and motion quality; (5) Generative audio — ElevenLabs partnership enabling AI-generated sound effects, music, and voice in Premiere. Skeptical perspectives: Adobe''s track record at MAX shows many announced features never reaching shipping software; the "creativity automation" framing raises concerns that AI tools designed to eliminate creative labor will reduce the economic foundation supporting the professional creative community that uses Adobe''s products; the multi-model marketplace approach (choosing between Google, Black Forest Lab models) introduces complexity and raises questions about consistent IP protection across training datasets of different partners. Adobe''s strategic position: positioned as the "safe enterprise AI" for commercial use through Firefly''s licensed training data, but the multi-model partnership strategy dilutes this advantage if third-party model IP protections are less rigorous.
