Artificial intelligence and evidence for social work: will a robot steal your job?
Bibliographic record
Abstract
Artificial intelligence (AI) is widely used to support decision making and interventions, arguably saving time, reducing bias and improving decision accuracy. The profession must urgently appraise the potential and pitfalls of this rapidly developing technology. This challenge was addressed at the 2024 European Social Work Research Association conference in Vilnius, at which the Evidence into Practice Special Interest Group focused on three contemporary AI developments: (1) large language models (LLMs); (2) AI- and robot-supported interventions; and (3) predictive risk modelling (PRM). This short ‘Reflection, exchange and dialogue’ article outlines the presentations, issues discussed and further reflections. Although LLMs have an impressive ability to manipulate language, essential case detail and analysis remain human tasks. There are robot technologies already helping people in the domains of disability and eldercare, and AI ‘language robots’ are being used favourably in low-risk mental health contexts, providing a non-judgemental (non-human) and ever-available ‘listener’ and ‘advisor’. PRMs raise many conflicting views. The ‘black box’ of AI may ‘hide’ systemic bias, though proponents argue that humans are biased too, so perfection is not an appropriate comparator. Our conclusion is that a priority is to examine, shape and regulate the interface between humans and computer algorithms.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.072 | 0.158 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.003 | 0.027 |
| Scholarly communication | 0.012 | 0.028 |
| Open science | 0.002 | 0.008 |
| Research integrity | 0.011 | 0.011 |
| Insufficient payload (model declined to judge) | 0.006 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".