A new article by Will Douglas Heaven, senior AI editor at the MIT technology review has called for an end to the use of predictive policing and justice, powered by AI algorithms. The article looks at a number of ways that race feeds into AI algorithms, and how this can detriment minorities. The article suggests that current AI systems, when applied to justice, end up continuing to reinforce existing systemic racism, and potentially lead to an increased bias, as judgement formed by a supposedly objective system, then reinforces existing bias.
Heaven, therefore, suggests that until AI has been developed to the point where it can be genuinely objective, it should not be used in such an important decision-making capacity, particularly as discussions continue in the US and globally around racism and bias in the justice system.
Visit the MIT technology review to read the full argument.
Comments on the lack of data available, provides a summary of a US household legal needs study in 1994 where 50% of households experienced one of more legal problems annually.
Of those with legal needs, 37% of the poor sought assistance from a third-party for resolution of the problem, 29% from a specifically legal third party such as a lawyer (21%) or other from a non-legal third party (8%). Among moderate-income households, assistance from a third-party was sought with 51% of problems, 39% from a specifically legal source (lawyers 28%, other legal/judicial 12%).
Comparing findings of US and UK legal need studies: the specific use of lawyers in the U.K. surveys is roughly the same as in the U.S.: 27% in England and Wales, 29% in Scotland versus 26% in the U.S. Where the substantial difference emerges is in the use of other third-parties. Moreover, because non-lawyers in the U.K. are authorized to give legal advice (such as volunteer-staffed Citizens Advice Bureaux or proprietary legal advice centres), the effective difference is even greater: Americans received advice from those who are able to give legal advice in only 37% of cases, compared to 60-65% of U.K. cases. Furthermore, a far smaller percentage of the U.K. respondents, as compared to U.S. respondents,
Gillian K. Hadfield. (2010) Higher Demand, Lower Supply? A Comparative Assessment of the Legal Landscape for Ordinary Americans Fordham Urban Law Journal.
Read article on Selected Works website
Should regulators intervene on the ground that it’s for the consumer’s own good, even if the consumer doesn’t know it?
In the Journal of Policy Analysis & Management, a series of articles puts the pros and cons of this approach, which seeks to intervene to address “costs we impose on ourselves by taking actions that are not in our own best interest.”
Access requires a subscription.
Mannix, B. F. and Dudley, S. E. (2015), THE LIMITS OF IRRATIONALITY AS A RATIONALE FOR REGULATION. J. Pol. Anal. Manage., 34: 705–712. doi: 10.1002/pam.21841