Skip to content

Tom Douglas’ Posts

Using AI to Predict Criminal Offending: What Makes it ‘Accurate’, and What Makes it ‘Ethical’.

Jonathan Pugh

Tom Douglas

 

The Durham Police force plans to use an artificial intelligence system to inform decisions about whether or not to keep a suspect in custody.

Developed using data collected by the force, The Harm Assessment Risk Tool (HART) has already undergone a 2 year trial period to monitor the accuracy of the tool. Over the trial period, predictions of low risk were accurate 98% of the time, whilst predictions of high risk were accurate 88% of the time, according to media reports. Whilst HART has not so far been used to inform custody sergeants’ decisions during this trial period, the police force now plans to take the system live.

Given the high stakes involved in the criminal justice system, and the way in which artificial intelligence is beginning to surpass human decision-making capabilities in a wide array of contexts, it is unsurprising that criminal justice authorities have sought to harness AI. However, the use of algorithmic decision-making in this context also raises ethical issues. In particular, some have been concerned about the potentially discriminatory nature of the algorithms employed by criminal justice authorities.

These issues are not new. In the past, offender risk assessment often relied heavily on psychiatrists’ judgements. However, partly due to concerns about inconsistency and poor accuracy, criminal justice authorities now already use algorithmic risk assessment tools. Based on studies of past offenders, these tools use forensic history, mental health diagnoses, demographic variables and other factors to produce a statistical assessment of re-offending risk.

Beyond concerns about discrimination, algorithmic risk assessment tools raise a wide range of ethical questions, as we have discussed with colleagues in the linked paper. Here we address one that it is particularly apposite with respect to HART: how should we balance the conflicting moral values at stake in deciding the kind of accuracy we want such tools to prioritise?

Read More »Using AI to Predict Criminal Offending: What Makes it ‘Accurate’, and What Makes it ‘Ethical’.

Asbestos neglect

  • by

Written by Tom Douglas

This is an unedited version of an article originally published by The Conversation

 

‘Calais Jungle Camp littered with asbestos’, ‘Buckingham Palace could be vacated to remove asbestos’, ‘Safety concerns for refugees and workers as Nauru asbestos removal program kicks off’.

Headlines such as these occur with monotonous regularity. Widespread asbestos use throughout much of the 20th century has ensured that the next contamination scandal is never far off, and asbestos-related legal decisions and personal tragedies often make the news as well. But despite the ongoing media attention, asbestos has not captured the public imagination as a public health threat, at least, not in comparison to other comparable threats like excessive sun exposure and drink driving.

Asbestos is a versatile fibrous mineral that can be cheaply mined and has unusual fire resistance and durability. Its use exploded in the twentieth century, when it was included in such diverse products as automobile brake linings, pipe insulation, ceiling and floor tiles, textured paints, concrete, mattresses, electric blankets, heaters, ironing boards and even piano felts. There is no safe threshold for exposure to asbestos dust, with even single exposures having been linked to cancer. Rates of asbestos-related cancer have recently been on the rise in Europe and Japan and look set to climb in many developing countries where asbestos is still being widely used, often without safety precautions. According to WHO estimates, asbestos now causes more deaths globally than excessive sun exposure. In the UK it is estimated to cause almost three times as many deaths as road traffic accidents.

Read More »Asbestos neglect

Engineering a Consensus:   Edit Embryos for Research, Not Reproduction

Written by Dr Chris Gyngell, Dr Tom Douglas and Professor Julian Savulescu

A crucial international summit on gene editing continues today in Washington DC. Organised by the US National Academy of Sciences, National Academy of Medicine, the Chinese Academy of Sciences, and the U.K.’s Royal Society, the summit promises to be a pivotal point in the history of the gene editing technologies.

Gene editing (GE) is a truly revolutionary technology, potentially allowing the genetic bases of life to be manipulated at will. It has already been used to create malaria-fighting mosquitoes, drought resistant wheat, hornless cows and cancer killing immune cells. All this despite the fact GE only become widely used in the past few years. The potential applications of GE in a decade are difficult to imagine. It may transform the food we eat, the animals we farm, and the way we battle disease.Read More »Engineering a Consensus:   Edit Embryos for Research, Not Reproduction

What’s the moral difference between ad blocking and piracy?

On 16 September Marco Arment, developer of Tumblr, Instapaper and Overcast, released a new iPhone and iPad app called Peace. It quickly shot to the top of the paid app charts, but Arment began to have moral qualms about the app, and its unexpected success, and two days after its release, he pulled it from the app store.

Why the qualms? For the full story, check out episode 136 of Arment’s excellent Accidental Tech Podcast and this blog post, but here’s my potted account: Peace is an ad blocker. It allows users to view webpages without advertisements. Similar software has been available for Macs and PCs for years (I use it to block some ads on my laptop), but Apple has only just made ad blockers possible on mobile devices, and Peace was one of a bunch of new apps to take advantage of this possibility. Although ad blockers help web surfers to avoid the considerable annoyance (and aesthetic unpleasantness) of webpage ads, they also come at a cost to content providers, potentially reducing their advertising revenue. According to Arment, the ethics of ad blocking is ‘complicated’, and although he still believes ad blockers should exist, and continues to use them, he thinks their downsides are serious enough that he wasn’t comfortable with being at the forefront of the ad blocking movement himself.

In explaining his reasons for withdrawing the app, Arment drew a parallel between ad blocking and piracy. He doesn’t claim that the analogy is perfect (in fact, he explicitly disavows this), and nor does he take it to be a knock-down objection to ad-blocking (presumably he believes that piracy is also morally complicated). But he does think there’s something to the comparison.

Like Arment, I think there are considerable moral similarities between ad blocking and piracy. But, also like Arment, ad blocking seems to me, intuitively, to be somewhat less morally problematic. This raises an obvious question: what’s the moral difference?

Read More »What’s the moral difference between ad blocking and piracy?

Should we punish crimes from the distant past?

Former Auschwitz SS officer Oskar Gröning is currently being tried as an accessory to murder for his role as an administrator in the extermination camp, and the trial has stirred up a lot of debate. One strand of the debate addresses the question whether Gröning was complicit in the extermination of prisoners, and whether he was culpable for this complicity. (Roger Crisp wrote a fascinating post on this a couple of weeks back.) But another strand – and the strand that I want to look at here – has addressed the question whether former Nazi war criminals should be tried and punished for deeds in their distant past. Eva Mozes Kor, an Auschwitz survivor and witness in Gröning’s trial has claimed that he shouldn’t be tried, though he should use his knowledge to help fight holocaust denial.

Let’s suppose that Gröning was indeed a culpable accomplice to murder. Should he then be punished? More generally, should serious crimes from decades go be punished? My intuition is that they should, but reflecting on why I have found it is not straightforward to defend this view.Read More »Should we punish crimes from the distant past?

Iterated in vitro reproduction and genetic orphans

In an article soon to be published in the Journal of Medical Ethics, Rob Sparrow imagines a procedure via which multiple generations of human embryos might be created in the laboratory. Egg and sperm cells would first be generated from existing or new human pluripotent stem cell lines. The resulting eggs would be fertilised using the sperm to create zygotes and ultimately embryos. Embryonic stem cells would then be harvested from these embryos and used to create new egg and sperm cells, which would in turn be used to fertilise one another to create further embryos. This process could be iterated, in principle indefinitely.

Let’s call this procedure ‘iterated in vitro reproduction’ (Sparrow calls it ‘in vitro eugenics’). Iterated in vitro reproduction is not yet possible, but, citing recent developments in the science of stem cell-derived gametes, Sparrow argues that it may well become so, though he acknowledges are number of significant hurdles to its development. He also discusses a number of possible applications of the technology and calls for an ethical debate on these. The most controversial application would be in the creation of designer children. Consider the following case, which is a variant on one of the scenarios imagined by Sparrow:

Jack and Jill present to a fertility clinic. Jack provides a sperm sample, and fertility doctors harvest a number of eggs from Jill. These eggs are fertilized with Jack’s sperm to create embryos, from which embryonic stem cells are derived. These stem cells are then induced to develop into eggs or sperm which are used to fertilise one another, and so on. The process is iterated numerous times, and at each stage, the embryos are genetically screened via pre-implantation genetic diagnosis. This screening is used to inform a process of selective crossing, so that, over several generations, the population of embryos evolves towards certain genetic dispositions desired by Jack and Jill (a disposition towards longevity, say). This process is aided by adding small amounts of genetic material from stem cell lines derived from other individuals. Eventually, doctors identify an embryo with almost exactly the desired combination of genes, and this embryo is implanted into Jill’s womb and carried to term. A child, Jarvis, is born.

Cases like this raise numerous ethical issues, some of which are discussed by Sparrow and the seven commentators on his paper. However, they also raise an interesting conceptual question: would the users of such a technology be the genetic parents of the resulting offspring? Would Jack and Jill be the genetic parents of Jarvis?

Read More »Iterated in vitro reproduction and genetic orphans

Destroying one ecosystem and constructing another: biodiversity offsetting and particular value

Suppose that the government is proposing a new policy regarding buildings of historical significance. Rather than simply banning the destruction of ‘listed’ buildings, the new policy would allow their destruction, provided that whoever destroys the building agrees to construct, somewhere nearby, a new building of a similar size, in a similar style, exhibiting a similar range of architectural innovations, and of a similar level of beauty. Blenheim Palace could be flattened and built over with a shopping mall and carpark, provided that mall developers agreed to construct a replica of the palace somewhere nearby.

Most would be disturbed by such a policy. Part of the reason that they would be disturbed, I presume, is that it seems to manifest a failure to recognise the true value of historical buildings. Not all of the value of historical buildings consists in their possession of generic properties like ‘being beautiful’, ‘being in the baroque style’ or ‘using space to dramatic effect’. Some of their value is value that they have as particular objects, and that could thus not be realised in any other object. Part of the value of Blenheim Palace derives from it’s being the birthplace of Winston Churchill. This value could not be realised in a replica of the palace built 5 miles down the road.

Of course, no-one is proposing a policy of sort I’ve just outlined. I bring it up because I think reflecting on this kind of case may throw some light on recent discussion regarding biodiversity offsetting (see, for example, here, here  and here).Read More »Destroying one ecosystem and constructing another: biodiversity offsetting and particular value

Living near a busy road can kill you

Early April saw some unusually smoggy days across much of Western Europe, resulting in widespread media attention to air pollution.
(See, for example, here, here and here.) On one day, air quality in some parts of London was worse than in Beijing. Further attention has been drawn to the issue by a number of recent official reports, including one from the World Health Organisation, which has declared that air pollution is now the world’s biggest single environmental threat to health.

As has been noted, media coverage can give a misleading picture of the health risks of air pollution. Coverage tends to focus on short-term peaks, such as those seen recently in Western Europe, but the health risks of air pollution are primarily related to long term exposure, and show no ’safe threshold’ effect. Elevated baseline levels of pollution are thus more of a problem than occasional peaks.

There’s another important aspect of air pollution that often goes unnoticed; small geographical differences can have a marked effect on exposure to air pollution and thus on risk of adverse health effects. For example, living near a busy road appears to substantially increase air pollution-related mortality. A study published last year in the Lancet (press summary here) investigated the effects of very local differences in air quality on mortality by pooling 22 European cohort studies. The investigators found that an increase in average annual fine particulate (PM2.5) exposure of 5 µg/m3 was associated with a 7% increase in the risk of dying from all natural causes. This is approximately the difference between living on a busy urban road and living in a traffic-free area. The finding was robust in the face of correction for various possible confounding factors.Read More »Living near a busy road can kill you

What’s wrong with adopting out an adopted child?

‘Re-homing’ is a term coined to describe the adopting out of adopted children. Reuters today published a long story on the practice, describing several cases in which children adopted from overseas by American parents were then put up for adoption again within the States, in one case only a few days after the initial adoption.

The adoptions described in the Reuters piece are problematic for obvious reasons. It appears that lax oversight arrangements in the US, particularly for intra-state adoptions, have allowed couples to adopt out their children with little or no vetting of the new parents, in some cases into abusive families, and often without much consideration of the child’s interests. But might there also be less obvious considerations in play here?Read More »What’s wrong with adopting out an adopted child?

Treating ADHD may reduce criminality

Pharmaceutical treatment of attention deficit-hyperactivity disorder (ADHD) is associated with reduced criminality according to a study published yesterday in the New England Journal of Medicine. The study of over 25,000 Swedish adults with the disorder found that men undergoing pharmaceutical treatments had a 51% chance of committing at least one crime in a 4-year period compared to 63% for those not in treatment. The risk of criminality for women with ADHD was 25% for those taking medication, and 31% for those not in treatment. It’s possible, of course, that the reduction in criminality associated with treatment was due not to the treatment itself, but to other factors, such as desire to improve behaviour, which could have both motivated treatment and reduced criminality. However, even when the investigators adjusted for likely confounders, they found that treatment was associated with significantly reduced criminal offending. Thus, their findings are at least suggestive of a causal relationship between medication and reduced crime.

It will be interesting to see how such a relationship, if it can be further supported, will be viewed by the general public and medical profession. Will it be seen as strengthening or weakening the case for ADHD treatment?

Read More »Treating ADHD may reduce criminality