
Lecturer
- About
-
- Email Address
- k.allan@abdn.ac.uk
- Telephone Number
- +44 (0)1224 273932
- Office Address
School of Psychology William Guild Building Room T6 Kings College Old Aberdeen AB24 3FX
- School/Department
- School of Psychology
Biography
I'm an experimental psychologist, and my research is now focussed on the psychology of human-AI interaction. We are trying to build and control Agents that perform specific tasks in stable ways underpinned by a 'Psychology'.
In the past my research has primarily addressed two questions. How does the brain manage to rapidly construct and update a reasonably good model of the world within our conscious experience? And, how does our model of the world avoid distortion or false beliefs when exposed to social influences from other people or from persuasive new technologies, like AI?
My background and training is in the cognitive neuroscience of long-term memory, stemming from my St Andrews University PhD work, using EEG to identify neural correlates of conscious and unconscious retrieval processes. I continued along similar lines, using PET and fMRI, in London as a research fellow at the UCL Institute of Cognitive Neuroscience, before moving to Aberdeen at the turn of the century. Here, I've continued to work on the neural basis of retrieval processing, but over the years I've grown increasingly interested in how our model of the world adapts to social influences that carry useful information or misinformation. In 2015, I began to work almost exclusively on cognitive and EEG-based diagnostics for Alzheimer's disease, leading to the formation of a spin-out company. In 2019, I returned to the School of Psychology full-time, where I've continued to research social influences upon cognition, in particular how the influence of AI may be psychologically regulated.
Internal Memberships
I am the School of Psychology's Lead for Online Course provision. I coordinate 4th year option courses (PS4040, Current Topics in Psychology, and PS4041 Critical Review) and one of our core MSc Conversion courses (PS5527).
- Research
-
Research Overview
Human-AI interaction - how do we engineer AI to be appropriately persuasive?
Adapting to social influences - how do we avoid incorporating other people's false beliefs or distorted views of the world into our own - obviously perfect - cognitive model of the world?
Research Areas
Accepting PhDs
I am currently accepting PhDs in Psychology, Computing Science.
Please get in touch if you would like to discuss your research ideas further.
Psychology
Accepting PhDsComputing Science
Accepting PhDsResearch Specialisms
- Artificial Intelligence
- Cognitive Neuroscience
- Cognitive Psychology
Our research specialisms are based on the Higher Education Classification of Subjects (HECoS) which is HESA open data, published under the Creative Commons Attribution 4.0 International licence.
Current Research
We are developing experimental protocols that allow us to control AI agent behaviour in stable ways to perform specific tasks. This allows us to then study how such AI interact with and, for example, persuade, humans to take their advice (or not).
We have specific projects running on human-AI interaction that investigate the propagation of bias (e.g. gender-bias) from AI into human decision-making, and how signals of accuracy, likelihood or confidence provided by AI influence trust in what they recommend.
Amplification of human gender stereotypical bias by advice from GPT models
There is huge societal concern about biased AI amplifying human biases. We have developed an experimental approach allowing biases to be controllably 'injected' into AI, e.g. Open AI's GPT models, to measure their influence relative to human baseline bias. We find that gender stereotypical biases injected into GPT's descriptions of people significantly amplify human baseline bias within 1st impressions.
Preprint of the in press paper: https://doi.org/10.31234/osf.io/r7vf5_v2
Collaborations
We have ongoing collaborations with Dr Gowri Sripada and Dr Georgios Leontidis from the School of Natural and Computing Sciences.
Supervision
I am currently supervising Mr Jacobo Azcona's PhD, which began in autumn 2022, looking at bias propagating from natural language AI to their human users, especially stereotypical biases that fuel gender discrimination.
Over the last 20 years, I have supervised or jointly supervised 11 PhD students, most recently Mr Lip Jin Tee (May, 2019, 'Potential biomarkers for early identification of individuals at risk of developing Alzheimer's disease') and Ms Maria Bulmer (Feb, 2020, 'The roles of episodic memory and semantic knowledge in individuation and stereotyping').
Funding and Grants
Currently, the lab has funding from the ESRC supporting PhD studentships (one began in Autumn 2022, and the other begins in Autumn 2025). Both are investigating human-AI interaction.
Over the last 20 years, I've received more than £1m in funding from various sources (including TauRx pharmaceuticals, ESRC, BBSRC, Bial Foundation, Carnegie Trust, SINAPSE) to investigate cognitive and EEG based diagnostics in Alzheimer's disease, the cognitive and neural basis of episodic memory, neurophysiological markers of suceptibility to memory distortion, and other topics.
- Teaching
-
Teaching Responsibilities
I currently teach - and coordinate - two 4th year option courses (PS4040/PS5040 Current Topics in Psychology, PS4041 Critical Review), as well as the 4th year Cognitive Neuroscience option course (PS4510/PS3524). I supervise 4th year honours projects, and I also teach on 3rd year methods courses.
Non-course Teaching Responsibilities
I lead the School of Psychology's provision of Online Courses, I sit on the School's Education committee, and I am a personal tutor.
- Publications
-
Page 1 of 5 Results 1 to 10 of 43
Stereotypical Bias Amplification, and Reversal, in an Experimental Model of Human Interaction with Generative AI
Royal Society Open ScienceContributions to Journals: Articles- [ONLINE] DOI: https://doi.org/10.1098/rsos.241472
Response to UKIPO Consultation on Copyright and Artificial Intelligence
11 pages.Other Contributions: Other ContributionsWorking memory capacity and self-cues: Consistent benefits in children and adults
British Journal of PsychologyContributions to Journals: ArticlesIn Search of a Goldilocks Zone for Credible AI
Scientific Reports, vol. 11, 13687Contributions to Journals: ArticlesRapid and reversible impairment of episodic memory by a high-fat diet in mice
Scientific Reports, vol. 8, 11976Contributions to Journals: ArticlesEffects of Sexually Dimorphic Shape Cues on Neurophysiological Correlates of Women's Face Processing
Adaptive Human Behavior and Physiology, vol. 3, no. 4, pp. 337-350Contributions to Journals: Articles- [ONLINE] DOI: https://doi.org/10.1007/s40750-017-0072-1
Simulation-Based Mentalizing Generates a “proxy” Self-Reference Effect in Memory
Quarterly Journal of Experimental Psychology, vol. 70, no. 6, pp. 1074-1084Contributions to Journals: ArticlesOld-new ERP effects and remote memories: the late parietal effect is absent as recollection fails whereas the early mid-frontal effect persists as familiarity is retained
Frontiers in Neuroscience, vol. 9, 00532Contributions to Journals: Articles- [ONLINE] DOI: https://doi.org/10.3389/fnhum.2015.00532
- [OPEN ACCESS] http://aura.abdn.ac.uk/bitstream/2164/5274/1/fnhum_09_00532.pdf
Breastfeeding and introduction of complementary foods during infancy in relation to the risk of asthma and atopic diseases up to 10 years
Clinical & experimental allergy, vol. 43, no. 11, pp. 1263-1273Contributions to Journals: Articles- [ONLINE] DOI: https://doi.org/10.1111/cea.12180
An obesogenic bias in women's spatial memory for high calorie snack food
Appetite, vol. 67, pp. 99-104Contributions to Journals: Articles- [ONLINE] DOI: https://doi.org/10.1016/j.appet.2013.03.011