VITAL STATISTICS. WHAT SHOULD EVERYONE KNOW ABOUT DATA?

Transcription

1 VITAL STATISTICS. WHAT SHOULD EVERYONE KNOW ABOUT DATA? Callum Rai 1807 WORDS

2 As we have progressed through the 21 st century the use of data has increased exponentially, with the amount of data on Earth set to grow by 10 times by However, with this meteoric growth, regulation has straggled behind leaving the public ill-informed about the fundamental changes within the statistical field and how it will impact them in the rapidly oncoming information era. The first section of this report will focus upon the uses of data within machine learning, the second portion will analyse the use of data and statistics within the media and in the final section the companies that possess large amounts of data will be examined. These three criteria will be examined on the basis of their potential positive and negative impacts to society. This evidence will give way to a verdict upon three basic principles everyone should know about data. Data in Machine Learning One of the most prominent uses of data within the modern day is in machine learning. Machine learning is the process by which computers learn, using massive volumes of Big Data 2. Today it has many real-world applications from driverless cars to chess AI. However, with around $30bn being spent annually on machine learning research 3 it is clear that it will form the foundations of many future developments within digital technology. But what impact will machine learning have upon society? One of the primary sectors machine learning is being used within is healthcare. It is widely believed that computers can diagnose illnesses more accurately and cost efficiently than a doctor using previous medical data. Already machine learning has shown its potential to do this with Google successfully producing a machine that can diagnose diabetic retinopathy as well as a medical professional 4. This technology can be replicated in economically deprived areas and reduce disease mortality with earlier diagnoses at a reduced price. There are many additional uses of this data-based technology within the medical industry, perhaps its most significant use in disease outbreak prevention. By using statistics upon previous epidemics, a computer can accurately predict potential disease outbreak using data such as temperature, rainfall and 1 David Reinsel, et al. (2017, April). The Evolution of Data to Life-Critical. Retrieved from (page 2) [Accessed 12th Feb. 2018] 2 SAP. (2017, February 15). What is Machine Learning? Retrieved from [Accessed 15th Feb. 2018] 3 Jacques Bughin, et al. (2017, June). Artificial intelligence the next digital frontier? Retrieved from tificial%20intelligence%20can%20deliver%20real%20value%20to%20companies/mgi-artificial-intelligence- Discussion-paper.ashx (page 3) [Accessed 15th Feb. 2018] 4 Peng, L. (2016, November 29). Detecting diabetic eye disease with machine learning. Retrieved from Google Blog: [Accessed 15th Feb. 2018] 1

3 humidity 5. This will allow authorities to react pre-emptively, exemplifying another way machine learning can aid in saving lives. This reduced level of mortality and medical spending could revitalise struggling economies further demonstrating the positive economic and social impacts machine learning can bring. However, a deep implication of using machine learning is the potential A.I. singularity. This is an event where a computer learns so much from the data they receive they become potentially destructive to humanity with their unparalleled intelligence. Professor Stephen Hawking believes it could spell end of the human race 6. This accentuates the power of data when used in conjuncture with digital technology to extents that humans remain unaware. The A.I. singularity epitomizes the prospect for machine learning to have unforeseen, colossal drawbacks potentially transforming the society as we know it. Another field where machine learning has been seen to have a positive impact is the education sector. Teaching machines can be produced using previous education data. The benefits of using a machine rather than a teacher are endless; 1 to 1 hyper-personalised teaching leads to a more tailored education which allows students to receive optimal teaching for their needs as the machine learns using the students work as data. As a result, students could become highly educated at a much earlier age aiding the economy. Furthermore, in less-economically developed areas this may allow for easier and more efficient education, giving equal access to high quality education across the globe. With a more literate population these countries may start developing more urban environments leading to a potential economic revolution. These possible economic and societal benefits give further evidence to how machine learning can benefit society. An additional implication behind the use of machine learning is the unemployment that occurs as a result. With machinery able to fulfil many human jobs at a lower cost and with higher productivity it effectively renders a large portion of the population redundant. It was estimated that 47% of jobs in America were at risk due to automation 7. In comparison during the Great Depression, the one of the worst recession event in American history, the unemployment rate was 25% 8. This risk to jobs due to automation as a result of data manipulation has the potential to revolutionise the society. The capitalistic model would now be outdated due to mass unemployment however current governments have failed to introduce discussion into what new model would be introduced to reshape society and prevent a potential anarchic situation. Yet again humanity remains unprepared and naïve to the dangers of using such advanced technology. 5 Vijeta Sharma, et al. (2015, December). Malaria Outbreak Prediction Model Using Machine Learning. Retrieved from [Accessed 15th Feb. 2018] 6 BBC News. (2014, December 2). Stephen Hawking: 'AI could spell end of the human race'. Retrieved from Youtube: [Accessed 15th Feb. 2018] 7 Carl Frey, et al. (2017, September 17). How susceptible are jobs to computerisation. Retrieved from (page 33) [Accessed 15th Feb 2017] 8 Frank, R. H., & Bernanke, B. S. (2006). Principles of Macroeconomics 3rd. Boston: McGraw-Hill/Irwin. (page 98) 2

4 Weighing this up, while it is clear the use of data in machine learning has its advantages, this potential good for society is completely over-shadowed by the unforeseen, potentially apocalyptic consequences of this advanced technology with the danger of a total societal breakdown. Hence from this we can conclude that one basic principle that should be known about data is its prospective danger when used in conjunction with technology. Statistics and the media Recent years have given way to a period of political turmoil. Amongst this chaos the term fake news has arisen to describe the rising number of news stories based upon baseless or false statistics within journalism. However, how is the use of statistics within the media affecting society? The upsides of the use of statistics in journalism is clear. It adds vital evidence to news stories to ensure they are not untruthful or made up. In this aspect statistics play a vital role in society in spreading the truth amongst the population playing a fundamental role in the modern-day democracy. Unfortunately, journalism is not so simple. Increasingly statistics are being used out of context in order to support news stories in a rising trend known as fake news. Accused participants of this range as far as the Russian government to the CNN. In % of Americans said that fake news caused them a great deal of confusion 9. This emphasises the increasing quantity of statistics that are used in the media to further the respective organisations agenda. From this it can be seen that exploitation of false statistics is an extreme threat to society threatening to sway opinions on all matter of topics based upon misled information. With the uprising of this perilous movement it is becoming obvious that although statistics provide indispensable numerical evidence, the wide-ranging availability of sources of media is being exploited by obscuring the epistemological context of which statistics in news stories are based upon. Therefore, a clear principle that should be known about data is that within the media, statistics are often used incorrectly or out of context to support their means. it is important to consider the source of these statistics or investigate further on a fact-checking website. 9 Statista. (2016, December). Level of confusion caused by fake news in the United States as of December Retrieved from Statista: [Accessed 18th Feb 2017] 3

5 Data Possession The use of data by big companies has been an area of debate for a long time with regulation struggling to keep pace with the rapidly evolving uses of personal information. And with such as Facebook receiving around 600Tb of personal data daily 10 it is important to know the nature of the impact that data privacy has on society. Many companies claim that the use of data allows them to provide a more personalised and efficient experience. For example, Google claims that by collecting data it provides a faster searching feature that provides results relevant to you 11, many companies reiterate a similar rhetoric. From this perspective a clear benefit of companies possession of this data is that it allows services to be more efficient and effective giving an advantage to society. A common counter argument against this is that the collecting of this personal data breaches one of the most basic human rights, the right to privacy. It is commonly thought that the collecting of personal data by digital companies and governments intrudes on the lives of the people. This is an increasingly important issue with over 40% of Americans believing that encryption technology that means data cannot be accessed by the government should be legal 12. Some believe this lack of data privacy could change the way society operates out of fear of punishment and moreover may be the first step to a surveillance state akin to that seen in the dystopian novel This exemplifies that drawback of collecting of personal data, the infringement of privacy and the negative implications this has on society. In more recent times personal data has shown its use to society in legal proceedings. The police frequently use data obtained using surveillance to bring larger criminals and terrorist to justice. Furthermore, latterly data obtained from smartphones on location and search history is now being offered to the police to serve as evidence in court proceedings 13. Data has been used to quickly take terrorists and criminals off the streets before they can cause more harm giving an example to a further positive use of collecting personal data. 10 Pamela Vagata, et al. (2014, April 10). Scaling the Facebook data warehouse to 300PB. Retrieved from Facebook: [Accessed 15th Feb 2017] 11 Google. (n.d.). Your Data. Retrieved from Google: [Accessed 15th Feb 2017] 12 Statista. (2016, May). Share of adults in the United States who think the government should be able to access encrypted communications when investigating crimes as of May 2016, by age group. Retrieved from Statista: [Accessed 18th Feb 2017] 13 Winter, J. (2015, November 6). How Law Enforcement Can Use Google Timeline To Track Your Every Move. Retrieved from The Intercept: [Accessed 15th Feb 2017] 4

6 A further apprehension towards the use of personal data is in discrimination. With personal data more readily available it is now commonly being used in screening procedures prior to interviews. It is thought that some employers use data found on the internet to unfairly discriminate against certain candidate on undesirable traits that would not be known otherwise. The use of this data remains a grey area in data protection laws 14. Further examples of perceived discrimination can be found in the insurance sector. It is commonly believed many insurance company set premiums based upon a wide variety of data including statistics obtained from social media profiles. Admiral Insurance recently stated that they priced car insurance for first time drivers based upon social media profiles 15. This illustrates how personal data can be used to discriminate people based upon extraneous factors. Overall, from this evidence we can see that the possession of our data from big companies has varying degrees of morality and usefulness to society. From this we can conclude that another principle we should know about data is that our personal data is possessed by many organisations and can be used both for and against our favour. What everyone should know about data. From exploring a range of issues surrounding data three principles everyone should know about data have been established: There are many prospective dangers associated with data when used in conjuncture with machine learning technology. Data and statistics are frequently used inappropriately within the media, it is important to investigate these facts further from other sources. Personal data is possessed by a wide range of organisations and can be used in both beneficial and harmful manners. 14 EmployLaw. (n.d.). The Data Protection Act for Employers. Retrieved from EmployLaw: [Accessed 18th Feb 2017] 15 Ruddick, G. (2016, November 2). Admiral to price car insurance based on Facebook posts. Retrieved from The Guardian: [Accessed 18th Feb 2017] 5