Harvard's Cass Sunstein: Algorithms can correct human biases
A tool that can slowly build a better world.
- Algorithms help drive the modern world.
- Algorithms reflect human biases, but some — as Harvard's Cass Sunstein notes — can be built to help correct our biases.
- If you build the right algorithm, you might be able to help contribute to a better world.
Algorithms are part of the engine that drives the modern world. When you search for something on Google, you're relying on a search engine defined by a specific algorithm. When you see what you see on your news feed on Facebook, you're not looking at something that comes to you naturally; you're looking at something defined by a specific algorithm.
There's been pushback recently on the idea of the efficacy with which algorithms make our world easier (which is part of the way in which algorithms are discussed — that they make our world easier). Some of the pushback is philosophical. Some of the pushback comes from an immediately practical place.
The pushback from a practical place takes the form of an article appearing in October of this year noting that Amazon got rid of an AI recruiting tool it was using that didn't like women. Another article from ProPublica noted that the algorithm used to determine whether or not a criminal defendant in the United States was liable to re-offend was racially biased.
Part of the reason why some algorithms are having trouble is because there are numerous mathematical ways to define the concept of 'fair', nor is every system built with enough flexibility to account for all the different ways in which 'fairness' can be defined. Consider a way by which one system assessed potential child abuse in the Pittsburgh region, as flagged in an article in Nature: "And, for reasons that are still not clear, white children that the algorithm scored as at highest risk of maltreatment were less likely to be removed from their homes than were black children given the highest risk scores."
But that doesn't mean that there aren't positive things at work among all this — there are, and Harvard Kennedy School professor Cass Sunstein recently released a paper to testify to that fact, arguing that "algorithms can overcome the harmful effects of cognitive biases."That being said, it's worth noting a strange moment in the paper where Sunstein writes that "... the decisions of human judges, with respect to bail decisions, show neither disparate treatment nor disparate impact. As far as I am aware, there is no proof of either." There is apparent proof, as noted in an article published in the Quarterly Journal of Economics. In the piece, the authors note that "estimates from Miami and Philadelphia show that bail judges are racially biased against black defendants, with substantially more racial bias among both inexperienced and part-time judges."
The paper from the Quarterly Journal of Economics renders slightly problematic the assertion that an algorithm can simply make an already race-blind decision-making process ('no proof of either') more efficient based on the mere potential for criminality. It also renders slightly problematic the notion that an algorithm can more or less reproduce what a Judge considering bail produces but a little bit more efficiently — with less crime and the like.
But this doesn't necessarily occlude some of what Sunstein points out about the particular algorithm noted in a paper put out under the auspices of the National Bureau of Economic Research:
1. "Use of the algorithm could maintain the same detention rate now produced by human judges and reduce crime by up to 24.7 percent. Alternatively, use of the algorithm could maintain the current level of crime reduction and reduce jail rates by as much as 41.9 percent ... thousands of people could be released, pending trial, without adding to the crime rate."
2. " … judges release 48.5 percent of the defendants judged by the algorithm to fall in the riskiest 1 percent. Those defendants fail to re-appear in court 56.3 percent of the time. They are rearrested at a rate of 62.7 percent. Judges show leniency to a population that is likely to commit crimes," treating "high-risk defendants as if they are low-risk when their current charge is relatively minor" while treating "low-risk people as if they are high-risk when their current charge is especially serious."
3. "If the algorithm is instructed to produce the same crime rate that judges currently achieve, it will jail 40.8 percent fewer African-Americans and 44.6 percent fewer Hispanics. It does this because it detains many fewer people, focused as it is on the riskiest defendants."
These are seemingly clear results that prove the thesis — that there is a bias that can be corrected to achieve a better result. Even with the complicating results found among inexperienced and part-time judges in Miami and Philadelphia, we can see that some judges interpret 'noise' as a signal. An algorithm can help provide a necessary aspect of clarity, and — potentially — justice.
While legalization has benefits, a new study suggests it may have one big drawback.
- A new study finds that rates of marijuana use and addiction have gone up in states that have recently legalized the drug.
- The problem was most severe for those over age of 26, with cases of addiction rising by a third.
- The findings complicate the debate around legalization.
Cannabis Use Disorder, is that when you get so high you can’t figure out how to smoke anymore?<div class="rm-shortcode" data-media_id="hfrVfwoH" data-player_id="FvQKszTI" data-rm-shortcode-id="0e62d9cb9c0a2361f81e9b5278706614"> <div id="botr_hfrVfwoH_FvQKszTI_div" class="jwplayer-media" data-jwplayer-video-src="https://content.jwplatform.com/players/hfrVfwoH-FvQKszTI.js"> <img src="https://cdn.jwplayer.com/thumbs/hfrVfwoH-1920.jpg" class="jwplayer-media-preview" /> </div> <script src="https://content.jwplatform.com/players/hfrVfwoH-FvQKszTI.js"></script> </div> <p><a href="https://www.ncbi.nlm.nih.gov/books/NBK538131/" target="_blank">Cannabis use disorder</a>, also known as CUD or cannabis/marijuana addiction, is a psychological disorder described in DSM 5 as "the continued use of cannabis despite clinically significant impairment." This includes people being unable to cut down on their usage despite wanting to, those who often use it despite finding it severely impairs their ability to function, or those who are putting themselves in danger to secure access to the drug.</p><p>While an understanding that marijuana can be addictive has existed for some time, and the image of the pothead who smokes so much they can hardly function is prevalent in our society, the effects of legalization on addiction rates have somehow gone understudied until now. Importantly, previous studies had failed to consider usage rates amongst populations over the age of 25.</p><p>In the new study, published in <a href="https://jamanetwork.com/journals/jamapsychiatry/article-abstract/2755276?utm_campaign=articlePDF%26utm_medium%3darticlePDFlink%26utm_source%3darticlePDF%26utm_content%3djamapsychiatry.2019.3254" target="_blank">JAMA Psychiatry</a>, focused on self-reported data on monthly drug use in four states where marijuana is now legal, Colorado, Washington, Alaska, and Oregon, from both before and after the drug was legalized in each state and compared it to others which have not yet legalized.</p><p>The data gave insights into the drug use habits of the respondents and specifically gave information about if they had smoked at all in the last month, the frequency of their drug use, and if they had ever had issues with how much they were using drugs.The researchers ultimately considered the responses of 505,796 individuals.</p><p>The increase in cannabis usage they found was <a href="https://www.newsweek.com/cannabis-use-disorder-rising-us-states-where-weed-legal-1471170" target="_blank">considerable</a>. The number of respondents over the age of 26 who claimed to have used the drug in the last month went up by 23% compared with their counterparts in states that have yet to legalize. Abuse of the drug by this group rose by 37%. </p><p>Teen usage rose by 25%, and addiction rates rose as well. This increase was small, though, and the authors have suggested it may be due to an unknown factor. The rate of usage or abuse for respondents between the ages of 18 and 25 did not increase at all. </p><p>After breaking the results down by demographics, the primary finding held; adults over the age of 26 are using marijuana more often when it is legalized, and they are starting to use it too much.</p>
The grain of salt<p>As in any study where findings are self-reported, the exact numbers you see here should be taken with a grain of salt. They could be slightly higher or lower. As this study relies on people self-reporting their usage of a drug that is still illegal in many places, it is very possible that the apparent spike in addiction rates is caused by more accurate reporting, as people who live in an area where pot is still illegal may be less likely to report smoking it every day.</p><p>And it should be repeated a thousand times over that correlation and causation are not the same thing. There could be some unknown factor causing these increases in each case. </p><p>Despite these qualifications, the study is still useful in giving us a general sense of what may happen in states that have yet to legalize. </p>
What does this mean for society and drug users?<div class="rm-shortcode" data-media_id="BdVRmwgX" data-player_id="FvQKszTI" data-rm-shortcode-id="d5c2f9e3739c26170f98b48bf07a3444"> <div id="botr_BdVRmwgX_FvQKszTI_div" class="jwplayer-media" data-jwplayer-video-src="https://content.jwplatform.com/players/BdVRmwgX-FvQKszTI.js"> <img src="https://cdn.jwplayer.com/thumbs/BdVRmwgX-1920.jpg" class="jwplayer-media-preview" /> </div> <script src="https://content.jwplatform.com/players/BdVRmwgX-FvQKszTI.js"></script> </div> <p>While claims of "reefer madness" are greatly exaggerated, marijuana has several well established and thoroughly studied side <a href="https://en.wikipedia.org/wiki/Long-term_effects_of_cannabis#Mental_health" target="_blank">effects</a>. While occasional use isn't terribly harmful, addiction can be. Lead author Magdalena Cerdá of New York University explains in the study that heavy marijuana use is associated with "psychological and physical health concerns, lower educational attainment, decline in social class, unemployment, and motor vehicle crashes."</p><p>A substantial increase in the number of people who are addicted to the stuff will incur costs to society down the line. <strong></strong></p><p>Of course, a 37% increase in problematic usage means that the percentage of adults smoking too much went from .9% to 1.23% of the population responding to the survey. This makes it far less prevalent than issues with alcohol, which affected around 6% of all Americans in <a href="https://www.verywellmind.com/prevalence-of-alcoholism-in-the-united-states-67876" target="_blank">2018</a>. </p><p>Recently, Big Think's <a href="https://bigthink.com/u/philip-perry" target="_self">Philip Perry</a> wrote a piece about how <a href="https://bigthink.com/want-to-protect-the-health-of-35-million-americans-legalize-cannabis" target="_self">legalization could improve the health of millions</a> by allowing the government to regulate the purity of commercially sold marijuana. This remains true. However, it must be weighed against the findings of this study, which suggests that at least some of these health gains will be wiped out by increased addiction rates.</p>
What does this mean for legalization efforts?<div class="rm-shortcode" data-media_id="bnPA9J9g" data-player_id="FvQKszTI" data-rm-shortcode-id="429e1d17ba031b02d4e79b4f02f54ab5"> <div id="botr_bnPA9J9g_FvQKszTI_div" class="jwplayer-media" data-jwplayer-video-src="https://content.jwplatform.com/players/bnPA9J9g-FvQKszTI.js"> <img src="https://cdn.jwplayer.com/thumbs/bnPA9J9g-1920.jpg" class="jwplayer-media-preview" /> </div> <script src="https://content.jwplatform.com/players/bnPA9J9g-FvQKszTI.js"></script> </div> <p>The legalization steamroller will undoubtedly keep rolling along. While health concerns are one factor in the debate over marijuana, it is only one of many. In Illinois, where I live, weed will become legal on January 1<sup>st</sup> of 2020. The legalization campaign and <a href="https://www.chicagotribune.com/politics/ct-met-illinois-recreational-marijuana-legislation-20190531-story.html" target="_blank">legislation</a> were more concerned with issues of social justice, the failures of prohibition, and finding a new source of tax revenue (<a href="https://www.chicagotribune.com/news/breaking/ct-illinois-tickets-collection-agencies-20190703-20190711-gyf77w52mbcdxkaxpleeay277a-story.html" target="_blank">since we're half broke</a>) than with matters of potential addiction.</p><p>As Vox <a href="https://www.vox.com/policy-and-politics/2019/11/13/20962924/marijuana-legalization-use-addiction-study" target="_blank">reports</a>, the authors of the study aren't suggesting that legalization shouldn't take place; that is another, broader debate. They merely wish to present the fact that legalization has a particular side effect that we should be aware of.</p><p>While this study is unlikely to change anybody's stance on if weed should be legalized or not, it does show us a critical element to be considered when discussing drug policy. No drug is perfectly safe, and we have reason to believe that legalizing marijuana will mean that more people will have a hard time with it. Let's hope that legalization proponents keep that in mind as they rack up their victories. </p>
Tea and coffee have known health benefits, but now we know they can work together.
Credit: NIKOLAY OSMACHKO from Pexels
- A new study finds drinking large amounts of coffee and tea lowers the risk of death in some adults by nearly two thirds.
- This is the first study to suggest the known benefits of these drinks are additive.
- The findings are great, but only directly apply to certain people.
Maybe you should enjoy this article with a cup of coffee or tea.<p> The <a href="https://drc.bmj.com/content/8/1/e001252?T=AU" target="_blank" rel="noopener noreferrer">study</a> involved 4,923 type 2 diabetics living in Japan. The average participant was 66 years old. All of the participants were taken from the rolls of the Fukuoka Diabetes Registry, a study geared at learning about the effects of new treatments and lifestyle changes on the health of diabetics. <br> <br> The participants filled out questionnaires concerning their health, diet, habits, and other factors. Among the questions were two focused on determining how much green tea or coffee, if any, the participants consumed over the course of a week. The health of the participants was recorded for five years. During this time, 309 of the test subjects died from a variety of causes. <br> <br> Subjects who drank more than one cup of tea or coffee per day demonstrated lower odds of dying than those who had none. Those who consumed the most tea and coffee, more than four and two cups a day, respectively, enjoyed the most significant reductions in their risk of death. This level of consumption was associated with a 40 percent lower risk of <a href="https://www.sciencedaily.com/releases/2020/10/201020190129.htm" target="_blank" rel="noopener noreferrer">death</a>. </p><p>Most interestingly, the effects of drinking tea and coffee appear to combine to reduce risk even further. Those who reported drinking two or three cups of tea a day and two or more cups of coffee were 51 percent less likely to die during the study, while those who drank a whopping four or more cups of tea and two or more cups of coffee had a 63 percent lower risk of <a href="https://www.medicalnewstoday.com/articles/diabetes-coffee-and-green-tea-might-reduce-death-risk" target="_blank" rel="noopener noreferrer">death</a>. </p>
So, should I start swimming in a vat of coffee and green tea?<iframe width="730" height="430" src="https://www.youtube.com/embed/LY0E-JQxeoY" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe><p> Not quite. </p><p> The primary takeaway from this study is that Japanese adults with type 2 diabetes who drink a lot of green tea and/or coffee die less often than similar people who do not. If this effect is caused by something in the drink, lifestyle choices people who drink that much tea all make, or something else remains unknown. The finding must be considered an association at this point. <br> <br> The eye-popping reductions in mortality rates are compared to the risk of death of others in the study. The people who died reported drinking less tea and coffee than those who lived. Unless you have several demographic and conditional similarities to the subjects of this study, you probably won't suddenly be at a two-thirds lower risk of death than your peers because you drink green tea. </p><p> Like all studies that depend on self-reporting, it is also possible that people misstated how much they consumed any one item. The study also did not look into other factors like socioeconomic status or education level, also known to impact death rates and potentially linked to coffee and tea consumption. </p><p> However, it is yet another study in the pile that suggests that <a href="https://www.healthline.com/nutrition/top-13-evidence-based-health-benefits-of-coffee" target="_blank" rel="noopener noreferrer">coffee</a> and <a href="https://www.healthline.com/nutrition/top-10-evidence-based-health-benefits-of-green-tea" target="_blank" rel="noopener noreferrer">green tea</a> are good for you. That much is increasingly <a href="https://www.health.harvard.edu/press_releases/health-benefits-linked-to-drinking-tea" target="_blank" rel="noopener noreferrer">agreed</a><a href="https://www.rush.edu/health-wellness/discover-health/health-benefits-coffee" target="_blank" rel="noopener noreferrer"> upon</a>. This study also suggests the benefits are additive, which is a new development.</p><p><br> So, while it isn't time to start the IV drip of green tea, a cup or two probably won't <a href="https://www.webmd.com/diabetes/news/20201022/coffee-green-tea-might-extend-life-for-folks-with-type-2-diabetes" target="_blank" rel="noopener noreferrer">hurt</a>. </p>
But most city dwellers weren't seeing the science — they were seeing something out of Blade Runner.
On Sept. 9, many West Coast residents looked out their windows and witnessed a post-apocalyptic landscape: silhouetted cars, buildings and people bathed in an overpowering orange light that looked like a jacked-up sunset.