Personalized uncertainty quantification in artificial intelligence

Tapabrata Chakraborti; Christopher R. S. Banerji; Ariane Marandon; Vicky Hellon; Robin Mitra; Brieuc Lehmann; Leandra Brauninger; Sarah McGough; Cagatay Turkay; Alejandro F. Frangi; Ginestra Bianconi; Weizi Li; Owen Rackham; Deepak Parashar; Chris Harbron; Ben MacArthur

doi:10.1038/s42256-025-01024-8

Personalized uncertainty quantification in artificial intelligence

Tapabrata Chakraborti, Christopher R. S. Banerji, Ariane Marandon, Vicky Hellon, Robin Mitra, Brieuc Lehmann, Leandra Brauninger, Sarah McGough, Cagatay Turkay, Alejandro F. Frangi, Ginestra Bianconi, Weizi Li, Owen Rackham, Deepak Parashar, Chris Harbron, Ben MacArthur

Research output: Contribution to journal › Review article › peer-review

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.

Original language	English
Pages (from-to)	522-530
Number of pages	9
Journal	Nature Machine Intelligence
Volume	7
Issue number	4
DOIs	https://doi.org/10.1038/s42256-025-01024-8
Publication status	Published - 23 Apr 2025

Access to Document

10.1038/s42256-025-01024-8

Cite this

Chakraborti, T., Banerji, C. R. S., Marandon, A., Hellon, V., Mitra, R., Lehmann, B., Brauninger, L., McGough, S., Turkay, C., Frangi, A. F., Bianconi, G., Li, W., Rackham, O., Parashar, D., Harbron, C., & MacArthur, B. (2025). Personalized uncertainty quantification in artificial intelligence. Nature Machine Intelligence, 7(4), 522-530. https://doi.org/10.1038/s42256-025-01024-8

@article{1e7f8d23f20a483e8116c50b32da3c1e,

title = "Personalized uncertainty quantification in artificial intelligence",

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.",

author = "Tapabrata Chakraborti and Banerji, \{Christopher R. S.\} and Ariane Marandon and Vicky Hellon and Robin Mitra and Brieuc Lehmann and Leandra Brauninger and Sarah McGough and Cagatay Turkay and Frangi, \{Alejandro F.\} and Ginestra Bianconi and Weizi Li and Owen Rackham and Deepak Parashar and Chris Harbron and Ben MacArthur",

year = "2025",

month = apr,

day = "23",

doi = "10.1038/s42256-025-01024-8",

language = "English",

volume = "7",

pages = "522--530",

journal = "Nature Machine Intelligence",

issn = "2522-5839",

publisher = "Springer Nature",

number = "4",

}

Chakraborti, T, Banerji, CRS, Marandon, A, Hellon, V, Mitra, R, Lehmann, B, Brauninger, L, McGough, S, Turkay, C, Frangi, AF, Bianconi, G, Li, W, Rackham, O, Parashar, D, Harbron, C & MacArthur, B 2025, 'Personalized uncertainty quantification in artificial intelligence', Nature Machine Intelligence, vol. 7, no. 4, pp. 522-530. https://doi.org/10.1038/s42256-025-01024-8

TY - JOUR

T1 - Personalized uncertainty quantification in artificial intelligence

AU - Chakraborti, Tapabrata

AU - Banerji, Christopher R. S.

AU - Marandon, Ariane

AU - Hellon, Vicky

AU - Mitra, Robin

AU - Lehmann, Brieuc

AU - Brauninger, Leandra

AU - McGough, Sarah

AU - Turkay, Cagatay

AU - Frangi, Alejandro F.

AU - Bianconi, Ginestra

AU - Li, Weizi

AU - Rackham, Owen

AU - Parashar, Deepak

AU - Harbron, Chris

AU - MacArthur, Ben

PY - 2025/4/23

Y1 - 2025/4/23

N2 - 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.

AB - 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.

UR - https://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=pure_starter&SrcAuth=WosAPI&KeyUT=WOS:001473586200002&DestLinkType=FullRecord&DestApp=WOS_CPL

U2 - 10.1038/s42256-025-01024-8

DO - 10.1038/s42256-025-01024-8

M3 - Review article

SN - 2522-5839

VL - 7

SP - 522

EP - 530

JO - Nature Machine Intelligence

JF - Nature Machine Intelligence

IS - 4

ER -

Personalized uncertainty quantification in artificial intelligence

Abstract

Access to Document

Other files and links

Fingerprint

Cite this