Personalized uncertainty quantification in artificial intelligence
Personalized uncertainty quantification in artificial intelligence
Artificial intelligence (AI) tools are increasingly being used to help make consequential decisions about individuals. While AI models may be accurate on average, they can simultaneously be highly uncertain about outcomes associated with specific individuals or groups of individuals. For high-stakes applications (such as healthcare and medicine, defence and security, banking and finance), AI decision-support systems must be able to make personalized assessments of uncertainty in a rigorous manner. However, the statistical frameworks needed to do so are currently incomplete. Here, we outline current approaches to personalized uncertainty quantification (PUQ) and define a set of grand challenges associated with the development and use of PUQ in a range of areas, including multimodal AI, explainable AI, generative AI and AI fairness.
522-530
Chakraborti, Tapabrata
26a5ab6f-fd15-4be2-bc8b-ed53f8913548
Banerji, Christopher r. s.
1f2450d6-5772-46b5-a913-2333f7b53a2a
Marandon, Ariane
5700cf99-720f-4044-b89e-2a43df252ffb
Hellon, Vicky
fac22100-0f7e-4a4b-8d6b-80f589feea77
Mitra, Robin
2b944cd7-5be8-4dd1-ab44-f8ada9a33405
Lehmann, Brieuc
5d1fc034-8a7f-4bc1-b475-b0ab1d86f914
Bräuninger, Leandra
5e9494a7-8709-4af2-9069-fbb5a016395d
Mcgough, Sarah
5e2ee4f8-dfbf-49c6-95ab-2bc63a8d14de
Turkay, Cagatay
fb6280da-9466-47be-839b-ca92491395d6
Frangi, Alejandro f.
dc8741d5-a350-4d3c-accf-ccf023a64b1d
Bianconi, Ginestra
654d0a71-cae6-4e5a-98c3-b40b3d8a2a96
Li, Weizi
efa499a5-fdc6-4e69-b4c1-d94b06af014d
Rackham, Owen
8122eb1f-6e9f-4da5-90e1-ce108ccbbcbf
Parashar, Deepak
57695253-ce8b-48f5-b7a1-97b416e0c889
Harbron, Chris
c9053c59-3f33-4842-aead-905f4a5b20ec
Macarthur, Ben
2c0476e7-5d3e-4064-81bb-104e8e88bb6b
23 April 2025
Chakraborti, Tapabrata
26a5ab6f-fd15-4be2-bc8b-ed53f8913548
Banerji, Christopher r. s.
1f2450d6-5772-46b5-a913-2333f7b53a2a
Marandon, Ariane
5700cf99-720f-4044-b89e-2a43df252ffb
Hellon, Vicky
fac22100-0f7e-4a4b-8d6b-80f589feea77
Mitra, Robin
2b944cd7-5be8-4dd1-ab44-f8ada9a33405
Lehmann, Brieuc
5d1fc034-8a7f-4bc1-b475-b0ab1d86f914
Bräuninger, Leandra
5e9494a7-8709-4af2-9069-fbb5a016395d
Mcgough, Sarah
5e2ee4f8-dfbf-49c6-95ab-2bc63a8d14de
Turkay, Cagatay
fb6280da-9466-47be-839b-ca92491395d6
Frangi, Alejandro f.
dc8741d5-a350-4d3c-accf-ccf023a64b1d
Bianconi, Ginestra
654d0a71-cae6-4e5a-98c3-b40b3d8a2a96
Li, Weizi
efa499a5-fdc6-4e69-b4c1-d94b06af014d
Rackham, Owen
8122eb1f-6e9f-4da5-90e1-ce108ccbbcbf
Parashar, Deepak
57695253-ce8b-48f5-b7a1-97b416e0c889
Harbron, Chris
c9053c59-3f33-4842-aead-905f4a5b20ec
Macarthur, Ben
2c0476e7-5d3e-4064-81bb-104e8e88bb6b
Chakraborti, Tapabrata, Banerji, Christopher r. s., Marandon, Ariane, Hellon, Vicky, Mitra, Robin, Lehmann, Brieuc, Bräuninger, Leandra, Mcgough, Sarah, Turkay, Cagatay, Frangi, Alejandro f., Bianconi, Ginestra, Li, Weizi, Rackham, Owen, Parashar, Deepak, Harbron, Chris and Macarthur, Ben
(2025)
Personalized uncertainty quantification in artificial intelligence.
Nature Machine Intelligence, 7 (4), , [14447].
(doi:10.1038/s42256-025-01024-8).
Abstract
Artificial intelligence (AI) tools are increasingly being used to help make consequential decisions about individuals. While AI models may be accurate on average, they can simultaneously be highly uncertain about outcomes associated with specific individuals or groups of individuals. For high-stakes applications (such as healthcare and medicine, defence and security, banking and finance), AI decision-support systems must be able to make personalized assessments of uncertainty in a rigorous manner. However, the statistical frameworks needed to do so are currently incomplete. Here, we outline current approaches to personalized uncertainty quantification (PUQ) and define a set of grand challenges associated with the development and use of PUQ in a range of areas, including multimodal AI, explainable AI, generative AI and AI fairness.
Text
Nature_MI__Revision
- Accepted Manuscript
Restricted to Repository staff only until 23 October 2025.
Request a copy
More information
Accepted/In Press date: 11 March 2025
e-pub ahead of print date: 23 April 2025
Published date: 23 April 2025
Additional Information:
Publisher Copyright:
© Springer Nature Limited 2025.
Identifiers
Local EPrints ID: 502019
URI: http://eprints.soton.ac.uk/id/eprint/502019
ISSN: 2522-5839
PURE UUID: 5990866a-9f4d-4d87-a1a6-17a43116cc2d
Catalogue record
Date deposited: 13 Jun 2025 16:50
Last modified: 14 Jun 2025 02:14
Export record
Altmetrics
Contributors
Author:
Tapabrata Chakraborti
Author:
Christopher r. s. Banerji
Author:
Ariane Marandon
Author:
Vicky Hellon
Author:
Brieuc Lehmann
Author:
Leandra Bräuninger
Author:
Sarah Mcgough
Author:
Cagatay Turkay
Author:
Alejandro f. Frangi
Author:
Ginestra Bianconi
Author:
Weizi Li
Author:
Deepak Parashar
Author:
Chris Harbron
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics
