Google figures out how to watermark AI-designed proteins
BiologyChemistryCybersecurity
THE AI ANGLE
Designing proteins while embedding cryptographic watermarksGoogle DeepMind developed SynthIDBio to watermark AI-designed proteins without altering their biological function. DNA synthesis providers can use cryptographic keys to confirm sequences came from trusted researchers. Unmarked designs can then receive closer scrutiny during threat screening.
THE TEACHING ANGLE
Watermark verification relies on statistical thresholds, but users can dilute signals by padding sequences or switching to other software.Read the original at arstechnica.com Generate teaching or study materials
More in Biology
- This light-powered AI can spot deepfakes with nearly 98% accuracyScienceDaily — AI · October 1, 2026
- AI Can Read Between The Redacted LinesAbove the Law · October 1, 2026
- Your Robot’s Safety Functions Already Work. What If the Input Lies?The Robot Report · October 1, 2026
- Mysterious oarfish's fins inspire large underwater robot designPhys.org — Biology · October 1, 2026
- AI scientist autonomously generates and validates new biological discoveriesPhys.org — Biology · October 1, 2026