ThinkFactory 2025: community discussion on harmonizing and accelerating self-driving laboratories

Thomas Pickles; Sterling G. Baird; Mohammad Salehian; Alastair J. Florence

doi:10.1016/j.matt.2026.102757

ThinkFactory 2025: community discussion on harmonizing and accelerating self-driving laboratories

Thomas Pickles^*, Sterling G. Baird, Mohammad Salehian, Alastair J. Florence

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

Abstract

ThinkFactory 2025 was a virtual workshop jointly hosted by the Acceleration Consortium and CMAC to facilitate discussion on harmonizing and accelerating self-driving laboratories. The three-hour event brought together more than 50 participants for plenary and breakout discussion across four themed tracks: AI and machine learning, data, orchestration, and robotics. Key priorities highlighted by the community included the need for clearer benchmarks to demonstrate value, improved accessibility for experimentalists without programming expertise, and stronger standards for interoperability, safety, and reproducibility. Participants also emphasised the importance of structured human–AI collaboration and shared infrastructure to promote openness and efficiency. This paper outlines the event structure and design and key discussion insights, providing a reproducible framework for hosting collaborative, cross-institutional virtual workshops.

Original language	English
Article number	102757
Journal	Matter
Early online date	31 Mar 2026
DOIs	https://doi.org/10.1016/j.matt.2026.102757
Publication status	E-pub ahead of print - 31 Mar 2026

Funding

This work was supported by the Canada First Research Excellence Fund (CFREF-2022-00042), through the Acceleration Consortium at the University of Toronto, and by the EPSRC MediForge Hub (Grant EP/Z532964/1), through CMAC at the University of Strathclyde.

Keywords

self-driving laboratories
accessibility
standards

Access to Document

10.1016/j.matt.2026.102757Licence: CC BY 4.0

Pickles-etal-Matter-2026-community-discussion-on-harmonizing-and-accelerating-self-driving-laboratoriesFinal published version, 3.41 MBLicence: CC BY 4.0

Cite this

@article{87127a9fbf344808aa94c5e3e282d800,

title = "ThinkFactory 2025: community discussion on harmonizing and accelerating self-driving laboratories",

abstract = "ThinkFactory 2025 was a virtual workshop jointly hosted by the Acceleration Consortium and CMAC to facilitate discussion on harmonizing and accelerating self-driving laboratories. The three-hour event brought together more than 50 participants for plenary and breakout discussion across four themed tracks: AI and machine learning, data, orchestration, and robotics. Key priorities highlighted by the community included the need for clearer benchmarks to demonstrate value, improved accessibility for experimentalists without programming expertise, and stronger standards for interoperability, safety, and reproducibility. Participants also emphasised the importance of structured human–AI collaboration and shared infrastructure to promote openness and efficiency. This paper outlines the event structure and design and key discussion insights, providing a reproducible framework for hosting collaborative, cross-institutional virtual workshops.",

keywords = "self-driving laboratories, accessibility, standards",

author = "Thomas Pickles and Baird, \{Sterling G.\} and Mohammad Salehian and Florence, \{Alastair J.\}",

year = "2026",

month = mar,

day = "31",

doi = "10.1016/j.matt.2026.102757",

language = "English",

journal = "Matter",

issn = "2590-2385",

publisher = "Cell Press",

}

TY - JOUR

T1 - ThinkFactory 2025

T2 - community discussion on harmonizing and accelerating self-driving laboratories

AU - Pickles, Thomas

AU - Baird, Sterling G.

AU - Salehian, Mohammad

AU - Florence, Alastair J.

PY - 2026/3/31

Y1 - 2026/3/31

N2 - ThinkFactory 2025 was a virtual workshop jointly hosted by the Acceleration Consortium and CMAC to facilitate discussion on harmonizing and accelerating self-driving laboratories. The three-hour event brought together more than 50 participants for plenary and breakout discussion across four themed tracks: AI and machine learning, data, orchestration, and robotics. Key priorities highlighted by the community included the need for clearer benchmarks to demonstrate value, improved accessibility for experimentalists without programming expertise, and stronger standards for interoperability, safety, and reproducibility. Participants also emphasised the importance of structured human–AI collaboration and shared infrastructure to promote openness and efficiency. This paper outlines the event structure and design and key discussion insights, providing a reproducible framework for hosting collaborative, cross-institutional virtual workshops.

AB - ThinkFactory 2025 was a virtual workshop jointly hosted by the Acceleration Consortium and CMAC to facilitate discussion on harmonizing and accelerating self-driving laboratories. The three-hour event brought together more than 50 participants for plenary and breakout discussion across four themed tracks: AI and machine learning, data, orchestration, and robotics. Key priorities highlighted by the community included the need for clearer benchmarks to demonstrate value, improved accessibility for experimentalists without programming expertise, and stronger standards for interoperability, safety, and reproducibility. Participants also emphasised the importance of structured human–AI collaboration and shared infrastructure to promote openness and efficiency. This paper outlines the event structure and design and key discussion insights, providing a reproducible framework for hosting collaborative, cross-institutional virtual workshops.

KW - self-driving laboratories

KW - accessibility

KW - standards

U2 - 10.1016/j.matt.2026.102757

DO - 10.1016/j.matt.2026.102757

M3 - Article

SN - 2590-2385

JO - Matter

JF - Matter

M1 - 102757

ER -

ThinkFactory 2025: community discussion on harmonizing and accelerating self-driving laboratories

Abstract

Funding

Keywords

Access to Document

Fingerprint

Cite this