Google’s latest AI arsenal rollout may feel like a symphony of innovation, but beneath the polished demos and preview tags, a more complicated melody plays. Announced during its Cloud Next event, these upgrades to the Vertex AI platform mark a deliberate escalation in Google’s bid to dominate enterprise generative tools—music, visuals, and now cloned voices included.

One of the most striking entrants is Lyria, a text-to-music model positioned as a rival to stock audio libraries. Users can compose full-length tracks in various genres with just a few prompts. But what feels like creative empowerment could also blur lines between originality and imitation—especially since developers actually train these models on unclear data sources.

Veo 2, Google’s developing video generator, now lets users remove logos, remove objects, stretch frames, and even simulate drone footage with AI-directed camera movements. These tools could revolutionize content creation, but they also raise stakes around deepfakes, visual manipulation, and the integrity of AI-made media—especially when companies like Google keep the training data behind closed doors.

And then there’s voice cloning. Google’s Instant Custom Voice, powered by its Chirp 3 model, promises multilingual speech synthesis with only 10 seconds of audio. A powerful tool, no doubt. But power, especially in tech, doesn’t come without risk. Google insists it has a review process to prevent people from abusing voices, but for many, the very idea that someone can clone their voice with a few seconds of tape feels chilling.

Notably, Google watermarks most generated content with SynthID. But with the exception of Chirp outputs, it’s unclear how enforceable or tamper-proof that safeguard truly is. Meanwhile, the company offers indemnity protections for users facing copyright flare-ups—but creators still left in the dark about whether their works were part of the training data may not find comfort there.

With this round of updates, Google isn’t just flexing its technological muscle—it’s shaping how creators will make, share, and use digital media. But in the rush to own the future, the present still holds unanswered questions.