You’re reading an issue of “The AI Economy,” my newsletter exploring the forces shaping the AI era—tracking how AI is rewriting business, work, technology, and culture. Subscribe to get expert insights and curated updates delivered straight to your inbox.
Sound is usually the last thing a creator adds to a project, and the first thing anyone watching it notices. It’s also the point where the work stops. The edit is finished, the pacing is right, and then comes the detour: out to a stock library, through a licensing page, into a subscription in search of a 30-second track. Adobe would prefer that creators never take that trip at all, which is why it’s making its music, speech, and sound-effects generation generally available in Firefly.
In other creative news, Adobe has introduced a free version of its Firefly AI Assistant. It’s also expanding its list of supported models in Firefly, adding Google’s Gemini Omni Flash to its portfolio.
Music, Speech, and Sound Effect Generation for All
Adobe first introduced its Firefly audio generation tools in 2025—sound effects in July and speech and music in October. All have been in beta until today. The Generate Music feature will create original, fully licensed tracks that fit a video’s length and mood. Generate Speech turns a script into clear and natural voiceovers powered by Adobe’s Firefly speech model or ElevenLabs. Generate Sound Effects creates custom sounds that match the action, timing, and energy of a scene.
Underscoring the significance of audio, Adobe cited a survey it conducted with the Berklee College of Music of video creators, musicians, and marketers. Nearly 80 percent of respondents (79.3 percent) said they posted video content daily or several times a week, and every respondent used music in their videos. Moreover, nearly a third (32.7 percent) revealed they’re already using AI-generated music in their final published work.
One reason why creatives would turn to Adobe Firefly for such generations is that the output would be commercially safe. It’s one of the distinguishing attributes of Adobe in this generative AI race—creatives, marketers, and others can rest easy knowing that the generations won’t run afoul of any copyright, legal, or licensing issues. Adobe said these are the top barriers cited by creators when using music.
“Using Generate Music, I am able to create a soundtrack that is unique to the video I’m making, and it gives me peace of mind because I know the music is fully licensed for commercial use,” creator Madeline Salazar said in a statement. “When I deliver something to a client, I’m putting my name behind the entire project, so being able to create original music for the piece, and feel confident about handing it over, is really important to me.”
Expanding Access to the Firefly AI Assistant
In April, Adobe launched the Firefly AI Assistant, a creative agent that connects to many Creative Cloud services. Instead of opening a single app and working within its constraints, Firefly handles it. Creatives just converse with the AI Assistant as they would with Claude or ChatGPT, describe what they want created, and the agent does what’s needed, orchestrating and executing complex, multi-step workflows across Photoshop, Firefly, Premiere, Express, Illustrator, and Lightroom.
Earlier this summer, Adobe enhanced its Firefly AI Assistant with new creative skills and tools. The agent can now create brand kits, short product videos, storyboards, and first-cut videos. There are also more ways for it to support additional workflows.
This assistant was initially limited to customers who purchased either a Creative Cloud Pro or a select Firefly plan (Pro, Pro Plus, or Premium). That’s since changed with the introduction of a “limited” free experience, giving those with a free Adobe account the ability to use the Firefly AI Assistant and an undisclosed number of daily generations.
Gemini Omni Flash Comes to Firefly
Adobe Firefly is more than a platform to showcase the company’s homegrown models to creatives. It’s to provide access to third-party models that mirror Adobe’s approach to AI: Providers agree not to use data provided by creators to train their models; models must be accessible through a single subscription; and all AI-generated assets must be digitally signed and include a verified certificate. Along with the native Firefly models, Adobe also supports models from Kling AI, Luma AI, OpenAI, Runway, Black Forest Labs, Ideogram, Moonvalley, Topaz Labs, Google, and others.
This week, Adobe added Gemini Omni Flash to that list, joining a laundry list of other Google models including Gemini 2.5 Flash, Gemini 3, Gemini 3.1, Imagen, Veo 2, Veo 3, and Veo 3.1 Fast. Introduced in June, Gemini Omni Flash is a multimodal model designed for high-end video and audio generation. Google said it performs well at generating video with synchronized audio. Outside of Firefly, it’s competitively priced at ten cents per second of video output.
