The Google DeepMind team has unveiled SynthID Bio, a watermarking technology tailored for biological uses. This innovative approach can subtly insert concealed watermarks into amino acid sequences of proteins generated by AI, as well as into three-dimensional protein structures predicted by AI, all without compromising their inherent biological functions. SynthID Bio encompasses two distinct watermark embedding techniques, one focusing on protein amino acid sequences and the other on three-dimensional structures. It guarantees that the proteins maintain their normal functionality and that structural predictions remain accurate, while also ensuring a high level of detectability for the watermarks. This technology can act as a valuable addition to current biosafety screening systems, offering provenance details for proteins designed by AI, aiding the DNA synthesis industry in refining its screening procedures, and upholding the data integrity of life science databases. Nonetheless, the technology currently grapples with the challenge of potential intentional tampering with the watermarks, and its resistance to such tampering must be bolstered in the future. The team is also exploring the application of this technology to the realm of genome creation.