Risk and eligibility assessments and facial recognition
- Specific benefits and concerns for each AI technology
- Risk and eligibility assessments and facial recognition
- Robotics
- LLMs and mental health chatbots
We asked about the following uses of assessing eligibility and risk using AI: to assess eligibility for welfare benefits, to predict the risk of cancer from a scan, to predict the risk of not repaying a loan, and facial recognition for policing and surveillance.
Across nearly all of these uses of AI, transparency in decision making also features as a commonly reported concern
Table 1: Most commonly selected benefits for risk and eligibility assessment technologies, and facial recognition
‘Which of the following, if any, are ways you think this technology will be beneficial?’
| Technology | Top three chosen benefits 2022/23 | Percentage |
|---|---|---|
| Assessing risk of cancer from a scan | 1 Earlier detection of cancer | 82% |
| 2 Less human error | 53% | |
| 3 More accurate than doctors | 42% | |
| Assessing loan repayment risk | 1 Faster and easier | 52% |
| 2 Less likely to discriminate | 39% | |
| 3 Less human error | 37% | |
| Assessing welfare eligibility | 1 Faster and easier | 43% |
| 2 Save money | 38% | |
| 3 Less human error | 37% | |
| Facial recognition for policing | 1 Faster | 77% |
| 2 More accurate than professionals | 55% | |
| 3 Save money | 46% | |
| Technology | Top three chosen benefits 2024/25 | Percentage |
| Assessing risk of cancer from a scan | 1 Earlier detection of cancer | 85% |
| 2 Less human error | 64% | |
| 3 More accurate than doctors | 46% | |
| Assessing loan repayment risk | 1 Faster and easier | 58% |
| 2 Less likely to discriminate | 44% | |
| 3 Less human error | 41% | |
| Assessing welfare eligibility | 1 Faster and easier | 52% |
| 2 Save money | 43% | |
| 3 Less human error | 39% | |
| Facial recognition for policing | 1 Faster | 89% |
| 2 More accurate than professionals | 66% | |
| 3 Save money | 51% |
As mentioned above, speed and accuracy commonly feature across perceived benefits for many AI technologies. For example, 85% of the UK public feel earlier detection of cancer is a potential benefit of AI tools that assess risk of cancer from a scan, and 66% of the public feel using facial recognition technologies in policing will be more accurate than police officers at identifying wanted criminals and missing persons. Alongside these, the public feel many AI tools may reduce mistakes made in carrying out the tasks we surveyed. This may be through reducing human error in decision making (e.g. 41% feel loan repayment risk tools will lead to less human error).
Concern around overreliance is most frequently reported for technologies that assess cancer risk (64%), technologies that assess welfare eligibility (60%), technologies that assess loan repayment risk (57%) and technologies that use facial recognition for policing (57%). Across nearly all of these uses of AI, transparency in decision making also features as a commonly reported concern.
Table 2: Most commonly selected concerns for risk and eligibility assessments and facial recognition
‘Which of the following, if any, are concerns that you have about the use of this technology?’
| Technology | Top three chosen concerns 2022/23 | Percentage |
|---|---|---|
| Assessing risk of cancer from a scan | 1 Overreliance on technology | 56% |
| 2 Accountability for mistakes | 47% | |
| 3 Transparency in decision-making | 32% | |
| Assessing loan repayment risk | 1 Accounting for individual differences | 52% |
| 2 Overreliance on technology | 51% | |
| 3 Transparency in decision-making | 49% | |
| Assessing welfare eligibility | 1 Overreliance on technology | 47% |
| 2 Accounting for individual differences | 55% | |
| 3 Accountability for mistakes | 47% | |
| Facial recognition for policing | 1 False accusations | 54% |
| 2 Accountability for mistakes | 48% | |
| 3 Overreliance on technology | 46% | |
| Technology | Top three chosen concerns 2024/25 | Percentage |
| Assessing risk of cancer from a scan | 1 Overreliance on technology | 64% |
| 2 Accountability for mistakes | 50% | |
| 3 Transparency in decision-making | 41% | |
| Assessing loan repayment risk | 1 Accounting for individual differences | 59% |
| 2 Overreliance on technology | 57% | |
| 3 Transparency in decision-making | 54% | |
| Assessing welfare eligibility | 1 Overreliance on technology | 60% |
| 2 Accounting for individual differences | 60% | |
| 3 Transparency in decision making | 54% | |
| Facial recognition for policing | 1 Overreliance on technology | 57% |
| 2 Technology will gather and share personal information | 56% | |
| 3 False accusations | 54% |
Image credit: Angelo Moleele on Unsplash