Risk Assessment for Implementing Watson at IVK
Any program that is introduced on such a scale as is being proposed for Watson to be used in IVK will come with a plethora of risks. To start, there are always internal cultural resistances to change that will need to be addressed. Additionally, proper training on the bias inherent in Artificial Intelligence (AI) will need to be of utmost importance to all who will be utilizing the program, or indeed the fruits of Watson’s labor (Chui, et al., 2018). Furthermore, with the increase in networking and digital connectedness , the threat of a digital breach in security increases exponentially. As with all services that base their servers elsewhere, the denial of services IVK will be utilizing can happen and will impact the day to day operations of the company, as well as the value that Watson will be adding. Legally, the risks to implementing an AI system into IVK is the idea of who will be held responsible if the machine reaches an incorrect conclusion or recommends a course of action that proves harmful. The biggest fear is that humans tend to believe that machines are inherently better at making decisions than humans, and will blindly trust the decision of an AI system that is specifically designed for the purpose (Violino, 2018). A major risk with Watson is the need for data sets to be cleaned and imported into the program in a manner in which the AI can efficiently produce information relevant to the company’s searches. This could result in massive over expenditure of capital as contractors would need to be hired to fill the immediate need and additional employees would need to be brought in and trained on the implementation of this new system.
Privacy and Data
IBM Watson has a few potential privacy issues worth noting. The first involves the customers. According to Private Internet Access Blog, “IBM acquired millions of medical records for their “Watson”. Granted, this IBM invested $150 million dollars building a new research center in Milan for its Watson Health division, which applies “cognitive computing” techniques to healthcare. In return for that $150 million investment, IBM will receive the medical records of 61 million Italians. Although the information that IBM reviecies is supposedly anonymous, it is said that medical records can never truly be anonymous (Moody, 2017).
AI as whole can truly be a privacy issue for customers, employee, vendors or anyone who uses it. According to Forbes article titled “Ethics And Artificial Intelligence With IBM Watson Rob High”, says that AI works best when it is transparent and humans and AI trust one another. With that being said, chat boxes are a very common use of AI and can also easily drill into one's personal information. For example, say I ask a chatbox what the balance is in my bank account because I want to invest money r make a large purchase. A futuristic chatbox could turn it into a more developed conversation and ask questions that drill deeper into one's personal information.Humans would need to feel comfortable supplying information to these machines. The article also says “the chances of AI truly going “rogue” are small, but they still need to be considered, and in order to maintain transparency and trust, the machine’s processes should be revealed (Morgan, 2017).”
IBM Watson prides itself for keeping data secure. On IBM’s website explaining Watson’s security they say “At IBM, we believe your data is yours – and yours alone. Therefore, it’s essential to create a system of best practices around data responsibility that guide the secure and ethical management of data...(IBM, 2017)”
Cybersecurity and Phishing
With that being said Watson seems to take their security measures very seriously. Some potential risks that could still find its way into encrypting Watson are by what are textbook describes as Phishing. Phishing “is a con executed using technology, typically targeted at acquiring sensitive information or tricking someone into installing malicious software (Gallaugher, 2017).” To help reduce the risk of phishing attacks, if IVK were using Watson they might want to have their employees learn more about how phishing attack happen. One of the EIA supplements can greatly help reduce that risk. Harvard Business Review posted as article titled, “I Was a Cyberthreat to My Company. Are You?“. The article explains how often phishing emails go out to employees and why its a huge cyber threat. There is a quiz attached to the article that all employee should take to see how much of a cyber threat they are. It could cost their company everything if they don't learn what is right and what wrong when clicking on software updates or links on emails (Prokesch, 2014). Employees should use this link to see how much of a cyber threat they are to their company:
Artificial Intelligence for businesses like IVK can have many risks and limitations. The initial capital investment of $480,000 is a huge risk factor for businesses looking to install a system like IBM Watson. It is not just the initial price that could cost IVK so much money. Any Artificial Intelligence has extra variable costs like the need for repairing and upgrading the system pretty frequently (Risks and limitations of artificial intelligence in business, 2018). Data is another major risk when implementing IBM Watson into IVK. Knowing what is good data and what is bad data is a major key when running an artificial intelligence system for a business. Many concerns can come into place when dealing with the data of IBM Watson; biased data, fake data, and the quality of your good data are all major concerns for IVK when they have to implement Watson into their business (Korolev, 2018). This then leads to another major limitation of AI systems. Installing the system into your business can take a lengthy amount of time. Going through the data and training the system and employees on how Watson will be used can be even more costly and time consuming.
References:
Chui, M., Manyika, J., Miremadi, M., Henke, N., Chung, R., Nel, P., & Malhotra, S. (2018, April). Notes from the AI frontier: Applications and value of deep learning. Retrieved from https://www.mckinsey.com/featured-insights/artificial-intelligence/notes-from-the-ai-frontier-applications-and-value-of-deep-learning#part4
Falkvinge, R. (2017, July 17). Millions of medical records for IBM's Watson: Not just a violation of privacy. Retrieved November 4, 2018, from https://www.privateinternetaccess.com/blog/2017/07/millions-medical-records-ibm-watson-multifaceted-privacy/
IBM. (2017, October 15). IBM Data Responsibility Perspective for Watson Data and AI*. Retrieved November 4, 2018, from https://www.ibm.com/watson/data-privacy/
John Gallaugher. (2017). Information Systems: A Manager’s Guide to Harnessing Technology. Boston, MA: FlatWorld
Korolev, M. (2018, February 13). AI’s biggest risk factor: Data gone wrong. Retrieved November 3, 2018, from CIO: https://www.cio.com/article/3254693/artificial-intelligence/ais-biggest-risk-factor-data-gone-wrong.html
Moody, G. (2017, May 22). Detailed medical records of 61 million Italian citizens to be given to IBM for its "cognitive computing" system Watson. Retrieved November 4, 2018, from https://www.privateinternetaccess.com/blog/2017/05/detailed-medical-records-61-million-italian-citizens-given-ibm-cognitive-computing-system-watson/
Morgan, B. (2017, June 13). Ethics And Artificial Intelligence With IBM Watson's Rob High. Retrieved November 4, 2018, from https://www.forbes.com/sites/blakemorgan/2017/06/13/ethics-and-artificial-intelligence-with-ibm-watsons-rob-high/#3cd54ab260e4
Prokesch, S. (2014, November 05). I Was a Cyberthreat to My Company. Are You? Retrieved November 4, 2018, from https://hbr.org/2014/08/i-was-a-cyberthreat-to-my-company-are-you
Risks and limitations of artificial intelligence in business. (2018, June 19). Retrieved November 3, 2018, from https://www.nibusinessinfo.co.uk/content/risks-and-limitations-artificial-intelligence-business
Violino, B. (2018, February 19). Risky AI Business: Navigating Regulatory and Legal Dangers to Come. Retrieved from https://www.cio.com/article/3256031/artificial-intelligence/risky-ai-business-navigating-regulatory-and-legal-dangers-to-come.html
No comments:
Post a Comment