Recent research has helped to cultivate growing awareness that machine-learning systems fueled by big data can create or exacerbate troubling disparities in society. Much of this research comes from outside of the practicing data science community, leaving its members with little concrete guidance to proactively address these concerns. This article introduces issues of discrimination to the data science community on its own terms. In it, we tour the familiar data-mining process while providing a taxonomy of common practices that have the potential to produce unintended discrimination. We also survey how discrimination is commonly measured, and suggest how familiar development processes can be augmented to mitigate systems' discriminatory potential. We advocate that data scientists should be intentional about modeling and reducing discriminatory outcomes. Without doing so, their efforts will result in perpetuating any systemic discrimination that may exist, but under a misleading veil of data-driven objectivity.

No abstract

Ann R Coll Surg Engl 2009; 91: 23-24 23Although the incidence seems to be increasing, testicular cancer remains an uncommon condition, with 1885 new cases occurring in the UK in 2003. 1 Public education has significantly increased the awareness of testicular cancer in men and emphasised the importance of regular testicular self-examination. 2 It may be a consequence of this education that secondary referral of men with scrotal abnormalities appears to be increasing. Although the majority of scrotal abnormalities are found to be benign, the discovery of a lump in the scrotum causes significant anxiety to men until they have the re-assurance of a urological opinion, perhaps supported by scrotal ultrasonography. Similar anxiety is suffered by men with testicular pain.In order to manage this increase in out-patient work-load, we set up a one-stop, rapid-access clinic for men with testicular lumps or pain. We present our first 30-month experience. Patients and MethodsWe established a clinic staffed by a consultant urologist and a specialist urological ultrasonagrapher. Clinics were held fortnightly with 25-30 men being seen each time. Three subgroups of patients were seen in the clinic:1. Men referred by general practitioners under the UK 'two-week rule' Department of Health guidelines 3 recommended for men with swellings in the body of the testis suspicious of malignancy.2. Men referred with other scrotal and testicular abnormalities.3. Men referred with a history of testicular or scrotal pain.The patients were first seen by the consultant urologist for history and clinical examination. Patients then underwent scrotal ultrasonography with a 7.5 MHz linear probe. The patient was then given an immediate diagnosis and treatment plan if necessary. Those patients with borderline intratesticular changes on ultrasound were followed up in the same clinic with a further testicular ultrasound at an appropriate interval. ResultsOver a 30-month period (January 2005 to July 2007), 845 new patients attended the one-stop testicular clinic and underwent clinical examination and scrotal ultrasonography. The median age of the patient was 26 years (mean, 41 years; range, 18-86 years). The majority (303 of 845; 33%) of patients were found to have no abnormality on clinical examination or testicular ultrasonography (Table 1). The most common abnormality found was an epididymal cyst (228 of 845; 27%). Only 33 of 845 (4%) patients had findings suspicious of testicular tumour and subsequently underwent radical orchidectomy. In two patients, ultrasound demonstrated testicular tumours (final pathology confirming malignancy) which clinical examination had failed to detect (Fig. 1).A total of 143 men were referred under the 'two-week rule' for suspected testicular cancer (i.e. thought by the referring practitioner to have a mass in the body of the testis). Of these, only 14 (10%) of men seen had sonographic findings suspicious for testicular cancer. PATIENTS AND METHODS We established a rapid-access testicular clinic staffed by a urologist and...

No abstract

This chapter takes up the issue of near-term artificial intelligence, or the algorithms that are already in place in a variety of public and private sectors, guiding decisions from advertising and to credit ratings to sentencing in the justice system. There is a pressing need to recognize and evaluate the ways that structural racism, sexism, classism, and ableism may be embedded in and amplified by these systems. The chapter proposes a framework for ethical analysis that can be used to facilitate more robust ethical reflection in AI development and implementation. It presents an ethical matrix that incorporates the language of data science as a tool that data scientists can build themselves in order to integrate ethical analysis into the design process, addressing the need for immediate analysis and accountability over the design and deployment of near-term AI.