Sample DATA4300 Data Security and Ethics Assignment

DATA4300 Data Security and Ethics Assignment Sample

Word count: 1500-2000
Weighting: 40 %

• you have been employed as a data ethics officer by an industry board (professional body) wanting to create a code of conduct around using artificial intelligence (ai) to (a) diagnose disease in a person earlier in the progression of this disease or (b) predict the spread of community diseases, in order to inform best practice in the healthcare industry.

• you are being asked to produce a framework for a code of conduct for a medical board.

• you can choose either of the two applications above (earlier disease diagnosis or community disease prediction using ai).

• this company code of conduct framework will also address individual responsibility as well as recommended government oversight.

• your framework will be presented in a report (a suggested structure is below).

Introduction

• introduction to the use of ai in medicine as a whole and fears related to its use, e.g. “seventy-one percent of americans surveyed by gallup in early 2018 believe ai will eliminate more healthcare jobs than it creates.”
source: https://healthitanalytics.com/news/arguing-the-pros-and-cons-of-artificial- intelligence-in-healthcare

• describe how ai is being used either to (a)diagnose disease in a person earlier in the progression of this disease or (b) predict the spread of community diseases
Data ethics issues

• outline possible data security, privacy and ethical issues associated with the use of patient data in ai. for example, why it may not be a good thing as stated in the quote below.

“it’s true that adding artificial intelligence to the mix will change the way patients interact with providers, providers interact with technology, and everyone interacts with data. and that isn’t always a good thing.”

source: https://healthitanalytics.com/news/arguing-the-pros-and-cons-of-artificial- intelligence-in-healthcare applicable principles

• outline theoretical and legal principles which are relevant to the data issues identified. afterall, if the algorithm gets it wrong who is to blame?

Solution

Introduction

Use of AI in medicine

Artificial Intelligence (AI) in electronic health records may be utilized for scientific research, increased efficiency, and medical management efficiency. Despite going through the traditional road of scientific publishing, recommendation formulation, and medical support tools, AI which has been properly constructed and taught with adequate data may assist in uncovering evidence-based medical practices from electronic health information. AI may also help design new patient care patterns of healthcare provision by studying clinical practice patterns obtained from electronic health data.

A critical feature of AI-based healthcare or medical research for assignment help is the utilization of data generated for electronic health records (EHR). If the underpinning system for information technology and network do not stop the dissemination of diverse or low-quality data, this data could be difficult to utilize.

Figure 1: AI in Medical
(Source: Walls 2022)

AI is being used in Predicting the spread of community diseases

In recent days, COVID-19 has been at its peak and even a kid is aware of its impacts and the disaster it caused to the world. For overcoming its impacts, AI is being used in several ways to the prediction of the spread of this communicable disease. As the major symptoms of any communicable disease are chest pain, cold, cough, etc. AI is being combined with other technologies for tracking and flagging possible carriers of the virus (Basu et al 2020).

AI-powered glasses for checking hundreds of individuals in minutes without establishing contact. This form of the monitoring system was employed at bus and railway terminals, in addition to other public areas with a high population density. accomplishing this by merging artificial intelligence with new temperature measuring technologies using computer vision. This method allowed for the contactless measurement of body temperature, a primary indicator of COVID-19, despite interfering with people's usual behavior. Anyone whose body temperatures surpassed the limit might be promptly identified using this technique. Since physical temp measuring is time-consuming and increases the danger of cross-infection due to the required interaction with others, it proved to be a successful solution (Basu et al 2020).

Figure 2: AI models
(Source: Piccialli et al. 2021)

• Through the use of a Support vector machine algorithm, AI separates out the data and identifies the disease spread (Agrebi and Larbi 2020).

• A combination of High-resolution accuracy and a Support vector machine leads to the identification and isolation of the disease with 100% accuracy.

• It identifies the patterns of data (signs) collected from the patients and through an algorithm AI cross-checks these signs with the right disease and leads to early prediction.

• Machine learning algorithms aid in detecting the red blood cell that got infected through malaria with the use of Digital in-line holographic microscopy data (Agrebi and Larbi 2020).

Figure 3: AI in predicting covid-19
(Source: Piccialli et al. 2021)

Data Ethics Issues

It is observed that for several years, researchers and other people have expressed worries about the ethical concerns of medical data storage and information security procedures, and Artificial Intelligence is increasingly dominating the discussion. Existing regulations are insufficient to safeguard an individual's personal health data.

Indeed, according to startling research, improvements in artificial intelligence have made the Health Insurance Portability and Accountability Act of 1996 (HIPAA) obsolete, and this was before the COVID-19 pandemic. The fact is that healthcare data is very significant to AI businesses, and numerous of them appear to not mind violating a few privacy and ethical standards, and COVID-19 has just worsened the situation. This is a huge concern for the protection of security seeing the hike in cybercrimes (Lexalytics 2021).

Figure 4: ethical and legal issues of AI in healthcare
(Source: Naik et al. 2022)

Below are mentioned the most probable data privacy, security, and ethical issues concerned with patient data in AI.

1. Continuously changing environment with regular disruptions - AI in healthcare must adapt to a constantly evolving environment with regular disturbances while adhering to ethical principles to safeguard patients' well-being. Nevertheless, a simple, crucial component of determining the security of any medical software is the ability to test the program and understand how the program might fail. The pharmacological and physiological processes of drugs or mechanical components, for example, are similar to the approach for software programs. ML-HCAs, on the other hand, maybe a "black box" problem, with workings that aren't apparent to assessors, physicians, or patients (Wanbil et al. 2018).

2. Uninformed consent in order to use patient’s data - Identity, reputation, and financial loss to the patient. When it comes to data of patients and people are more concerned about it and get in stress due to this, In the online services offered by the medical industry or health care industry tend to collect it's to collect a great piece of information about the patients to feed into the system which leads to data security issues. The information of the patients can be used to manipulate them In the future, creating fake identities, conducting cyber - crimes leading to financial and reputation loss, etc. All these concerns are the major ones that every patient is concerned about (Gupta et al. 2020).

3. Algorithmic biases and impropriety - No accountability for any harm done to patients as AI is a computerized system with no strict laws. Algorithms that may function by unwritten rules and develop new patterns of behavior which come under the AI are apparently threatening the ability to trace responsibility back to the developer or operator. The claimed "ever-widening" difference is the reason for concern since it affects "both the ethical structure of the community and the basis of the accountability concept in law." The adoption of AI may leave the healthcare industry and the patients with no one to hold responsible for any harm done. The scope of the threat is unclear, and the employment of technology will significantly restrict the human capacity to assign blame and accept responsibility for decision-making (Naik et al. 2022).

Figure 5: Cons of AI Adoption in Healthcare
(Source: Ilchenko 2020)

4. Lack of transparency and traceability in the system - The lack of computational and algorithm as well as operations of the computational system in AI collecting the patient’s information transparency has influenced many legal arguments about artificial intelligence. Because of the rising use of AI in elevated circumstances, there is a greater need for responsible, egalitarian, accessible, and transparent AI design and administration. The two most fundamental characteristics of visibility are data accessibility and comprehension. Information regarding algorithm functioning is usually purposefully made difficult to get (McKeon 2021).

5. Sourcing of data and Personal Privacy violation - With the World Data Corporation forecasting that the worldwide data sphere might very well develop from 33 zettabytes (33 trillion gigabytes) in 2018 to 175 zettabytes (175 trillion gigabytes) by the year 2025, businesses will have access to enormous amounts of both structured as well as unstructured data to mine, modify, and organize. As this data sphere expands at an accelerating rate, the dangers of revealing data owners or consumers including the patients and the staff of any organization and the hospital industry rise, and protection of personal privacy becomes more difficult to secure (McKeown 2022).

Figure 6: Annual data breach
(Source: McKeown 2022)

Whenever data leaks or breaches occur, the following repercussions may drastically harm an individual as well as indicate possible legal infractions, since many legislative bodies are increasingly enacting legislation that limits how personal information can be treated. The General Data Protection Regulation (GDPR) implemented by the European Union in April 2016 is a well-known regulation instance of this, which impacted the Consumer Privacy Act approved in June 2018 (McKeown 2022).

Applicable Principles

Legal principles against the AI issues in healthcare

• HIPAA requires regulated organizations to secure health information and patient records (information or data) when it relates to Protected Health Information (PHI). Dealing with any third-party provider has concerns that must be thoroughly examined. Whenever committing confidential material to an Artificial Intelligence vendor, healthcare institutions must create Business Associate Agreements (BAAs) to subject suppliers accountable for the same stringent data security requirements. As Artificial Intelligence technologies develop and healthcare businesses adopt AI into everyday activities, regulation loopholes remain to keep this technology in the shadows.

• Another principle that the healthcare industry must follow while implementing AI is that they need to offer full transparency to the system and the data information accountability of the patients in order to maintain their trust. In such cases, the placement and ownership of the computers and servers which keep and access patients' medical data for healthcare AI usage are critical. Without notable exceptions, laws must mandate that patient information be kept in the jurisdiction from where it was collected.

• Developing and applying artificial intelligence (AI) in order to strengthen national security & defense as well as strengthen the trusted collaborations by softening science and technology guidelines with the application of human judgment, particularly whenever an activity has the possibility to deprive people of civil liberties or intrude with one‘s fundamental freedoms of civil rights.

• Creating and implementing the best practices to increase the dependability, privacy, and precision of Artificial intelligence design, implementation, and usage. this will use best practices in cybersecurity to promote sustainable development and reduce the possibility of adversary impact.

• Legal principles shall offer adequate openness to the community and corporate clients about our AI methodologies, implementations, and uses, within the constraints of privacy, innovation, and reliability as defined by law and regulation, and in accordance with the IC's Principles of Information Visibility. it will create and implement systems to define roles and hold people accountable for the usage of AI and its results.

REFERENCES