Figure 1. Comparison of the workflow using the input of self-assembling pentapeptides from AI expert and human expert. The search range for peptides increases dramatically due to the multiple combinations of 20 amino acids, and 3.2 million pentapeptides are unmanageable to be calculated with the brute-force method. Human expert adopted rationally design methods such as hydrophobic scales, molecular patterns (npnpn), and personal experience. Six of the eleven pentapeptides proposed by six human experts were synthesised and observed to be clustered together. In contrast, AI expert developed a combination of MCTS. Six of the nine pentapeptides proposed by the AI expert were synthesised and found to be clustered. AI expert has also proposed novel sequences involving multiple amino acids to recover some intuitive sequences, reflecting its advantages in overcoming human bias. Atomic force microscope images of some promising pentapeptides from both groups are shown. Created with MedPeer (www.medpeer.cn). AI: artificial intelligence; AP: aggregation propensity; CG: coarse-gained; LC-MS: liquid chromatography-mass spectrometry; MCTS: Monte Carlo tree searches; MD: molecular dynamic; n: non-polar; p: polarity; RP-HPLC: reverse phase high performance liquid chromatography; SPPS: solid phase peptide synthesis.