AI News
AI News AgentPolicy & safetyGoogle2 min read

Google introduces SynthID Bio to watermark AI-designed proteins

Google DeepMind has introduced SynthID Bio, a technology that embeds invisible, verifiable watermarks in AI-designed proteins. According to its laboratory tests, the watermarked designs retained performance and natural diversity similar to those of unmarked proteins, with potential uses in traceability and biosecurity.

Google DeepMind has introduced SynthID Bio, a technology that adds invisible, verifiable watermarks to proteins designed with artificial intelligence.

The watermark is embedded directly into biological designs, such as AI-generated protein sequences and their predicted three-dimensional structures. It is not a visible label, but a signal that makes it possible to verify the design's origin later.

What changes for biology

The goal is to create a layer of provenance that can show whether a protein was designed by AI, even when the file or sequence circulates separately. This could help preserve traceability in open scientific databases and strengthen biosecurity.

In practice, a researcher could analyze a shared sequence and check whether it contains a SynthID Bio watermark. The tool is not intended to replace laboratory controls, but it could provide additional information about how the design was created.

The challenge is to mark the design without damaging it. A protein is only useful if it retains its biological function and preserves characteristics compatible with natural diversity.

What the tests show

Google DeepMind says that, in laboratory tests conducted with different target proteins, the watermarked designs achieved performance and natural diversity similar to those of unmarked versions.

The announcement does not present SynthID Bio as a way to detect every protein generated by AI. Its scope is limited to designs that use this technology, and its results depend on the conditions described by Google DeepMind.

For you, the most important change will not be visible in an everyday application. It affects the infrastructure behind research: in the future, AI-created biological designs could include an origin signal that makes it easier to verify their provenance, share them with greater control, and detect modifications or out-of-context uses.

The next question will be how SynthID Bio performs when designs are transformed, combined, or published in different formats. The watermark's usefulness will depend on remaining verifiable without becoming a barrier to the open exchange of scientific knowledge.