Polylog
← Trends

Provenance Marking for AI-Designed Biology

As generative protein design becomes routine, labs and regulators push for embedded provenance in the designed molecule itself rather than in the file, so expect recurring watermarking and attribution methods validated in wet-lab assays, arguments over function cost and removability, and biosecurity screening pipelines starting to assume marked designs.

weakening · confidence 27 · -7 7d · Emerging (watchlist) · tracking since October 1, 2026 · updated October 12, 2026

Sign in to get threshold and movement alerts for this trend.

Score history

Daily conviction score, 0 to 100. Higher means the thesis is more strongly corroborated.

Oct 11 · 28Oct 12 · 27

Now 27 · -1 since Oct 11 · ranged 27 to 28

Showing the last few days. Unlock full score history.

Why the conviction moved

  • Oct 1
    Strengthened

    DeepMind reported watermarking AI-designed proteins with no measurable loss of binding function, with wet-lab tests against VEGF-A, the SARS-CoV-2 spike receptor-binding domain and PD-L1 showing marked designs matching unmarked ones on hit rate and affinity.

Source trail

  • Supporting · October 1, 2026

    DeepMind Watermarks AI-Designed Proteins Without Measurable Loss of Binding Function

    DeepMind reported watermarking AI-designed proteins with no measurable loss of binding function, with wet-lab tests against VEGF-A, the SARS-CoV-2 spike receptor-binding domain and PD-L1 showing marked designs matching unmarked ones on hit rate and affinity.

    Google DeepMind

Unlock full source trail, score history, and daily updates.

Unlock Trends

Affected regions & assets

RegionsGlobal
Assets2 assetsUnlock Trends

Townsquare

Argue the thesis in Townsquare.