Mahboobin 4:00 R13 ETHICAL DILEMMA IN GRAPHENE-BASED PROCESSOR PRODUCTION Ronen Orland (roo18@pitt.edu) THE SITUATION As an engineer in the research and development department of AMD, I am working on creating commercially viable and manufacturable graphene-based CPUs, or central processing units, with the newly implemented graphene transistors. At this point in time, my lab has created a method to have the capability to synthesize graphene pure enough for electronics use and is able to do this at a large enough scale that mass-manufacturing is now a possibility. After much development and research, my team has been able to successfully implement the graphene transistors into the designed processor. With improved production methods, this processor can be made at a reasonable enough price that the general public can now purchase it (at this time, graphene-based processors are too expensive for the regular person to afford one, and are only sold to industry). However, after some further testing, one of my colleagues produces some unforeseen results. The stress tests show that the processors are prone to failure much sooner than regular processors already on the market. With regular use, their performance will substantially decrease after only a year. At that point, they are liable to fail and even short-circuit (excessive current flow through a circuit, causing damage [1]), not only ruining themselves but potentially other components in the computer. Yet another danger of the processor shortcircuiting is the possibility of it catching fire. Thus, not only does this scenario put a financial burden on the consumer, but also a safety hazard. The tests conclude that about 10% of all produced processors will fail in this way, and unfortunately, there is no way to determine which individual pieces will do so. The test results are reported to the department manager. After reviewing them, however, he decides to use the design to begin production of the graphene processors. The manager says that the company needs to put out the product as soon as possible to stay ahead of competitors, as there are no other graphene-based processors available to the general public as of yet. Being the first company to release this product would have immense benefits, such as high profits and a great amount of publicity. The manager claims that releasing the product as-is will, overall, do much more good than any possible harm. I voice my disagreement, and tell him that once the processors start to fail and cause damages, people will start to distrust both the technology and the company, but the manager is adamant on his decision and wants to move ahead regardless. CODES OF ETHICS University of Pittsburgh, Swanson School of Engineering 1 Submission Date 2015-11-03 As an engineer, I have codes of ethics that I must follow and uphold. All engineers must follow the code of ethics of the National Society of Professional Engineers (NSPE), and as a computer engineer I also follow the code of the Institute of Electrical and Electronics Engineers (IEEE). To make a decision on this matter, I must consult these codes and use them to determine what course of action is most ethical. NSPE Code of Ethics After consulting the code of the NSPE, I see that the very first fundamental principle is to “hold paramount the safety, health, and welfare of the public” [2]. As I am concerned for the well-being of the consumers and what could happen to them as a result of purchasing the processor, I look for guidance under this section. One of the subsections states that “Engineers shall not permit the use of their name or associate in business ventures with any person or firm that they believe is engaged in fraudulent or dishonest enterprise” [2]. This is a reasonable suggestion, as I do not want to be responsible for any harm that comes to others as a result of my research on the processor. If I do not condone going forward with the current design, I would not give my approval and thus not associate myself with the project any further. However, though this prevents me from being responsible for any harm that may be done, it does not prevent the damage from happening. Subsection A states that “If engineers’ judgement is overruled under circumstances that endanger life or property, they shall notify their employer or client and such other authority as may be appropriate” [2]. This statement applies fully to my situation, as I informed the manager of potential harm that may be done to consumers and their computers as a result of the current product design, but he ignored my warning and wanted to proceed with putting it into production. Based on this section of the code, I could go directly to a superior above the department manager and inform him/her fully of the situation and give my opinion on the matter. Going behind my superior’s back is not professional, but in this situation I must give precedence to the ethical dilemma. If I can convince a higher-up that this is the wrong course of action to take, I can uphold my ethical duty to put first the welfare of the public. IEEE Code of Ethics As a member of the IEEE, I also consult their code of ethics to see if it can assist me in determining what would be best for me to do. The very first statement of the code is almost identical to the aforementioned principle of the NSPE and states that I Ronen Orland must “accept responsibility in making decisions consistent with the safety, health, and welfare of the public, and to disclose promptly factors that might endanger the public or the environment” [3]. That I see this once again in another code of ethics reassures me that I must prevent the production of the processor until it is safe. The code also mentions acknowledging and correcting errors [3], which is vitally important in this case. Both the manager and I acknowledge the problem with the processor, but he does not wish to correct it. I can either convince him to delay production or figure out the problem before production even begins. Both options are difficult; convincing the manager would be difficult, as he has set his mind on one goal, while attempting to solve the problem in the product could range anywhere from several days to several long months. A very important principle that the code discusses is one to “assist colleagues and co-workers in their professional development and to support them in follow this code of ethics” [3]. If I could make the manager rethink his stance on the situation, I would not only be fulfilling my ethical duties by preventing harm from coming to the customers, but also by helping the manager realize that what he wants to do is wrong. incident. Whatever profits they made from the selling the faulty products would have been cut by having to replace them with new pieces. These exact same outcomes can happen to AMD if the manager decides to go through with selling our currently designed processor. The situation could even be worse, as our product’s defect would not just be an inconvenience, but a possible safety hazard. Lawsuits would almost assuredly be filed against the company once word gets out that we sell a product that may very well set itself ablaze. Overall, the possible profits to be made from selling the processor as it stands are simply not worth the consequences that will stem from the risk we place on the consumer. This case study assures me I must convince the manager to not go through with his plan. The Ford Pinto The second case study I look at is that of the Ford Pinto. Typically, the whole process of designing to finalizing a car’s design took Ford about three and a half years. The company decided that they didn’t have the time to spare in their current state and decided to cut the time down to two years, almost half the original. Before production could begin, prototypes of the model would have to be crash-tested and checked if they met the safety standards. The results showed that each prototype failed the 20-mph test, which states that an automobile must “withstand a rear-end impact of 20mph without fuel loss” [5]. In each case, the gas tank ruptured and caused dangerous leaks. Only prototypes modified specifically for the test passed. Ford eventually decided to pass the model for production after doing a monetary analysis of risk versus profit. In this analysis, they placed a value on a single person’s life and multiplied that by the estimated amount of deaths to result and compared that to the profits they would make. Because of their decision, customers with the Pinto who were rearended typically ended up dying from the typically non-lethal accident because the gas tank would rupture and explode. Not only did Ford knowingly produce a car that could cause harm, but they did so knowing that there would definitely be people who died as a result. This case is similar to my situation, but on a more extreme scale. There is a chance of my company’s product catching fire, but circumstances would have to be rather specific for it to spread and cause damage to beyond the computer it is housed in. Regardless, there is still the possibility of greater harm being caused, and in extreme cases of lives being put in danger. Unlike Ford, I do not place a monetary value on a person’s life, and I will most certainly not compare the value of potentially causing harm to a customer to the profit that can be made by selling them a deficient product. CASE STUDIES To further examine what I could and should do in my predicament, I look at several case studies. The Flaw in the Intel Pentium Chip The first case I read is that of the flawed Intel Pentium microprocessor in 1994. After release, the media found that the processor had a flaw. This discovery was significant since 80% of personal computers worldwide were using the Intel Pentium [4]. Evidently, flaws in complex components, such as processors, aren’t very rare. Typically, these flaws are very minor and go undetected the user and do not affect the machine. In this case, however, the flaw created incorrect answers when performing an operation that is frequently required by the processor. It suddenly became very clear to users that there was a major flaw in their Pentium chips. When confronted, Intel claimed that though “there was indeed a defect in the chip, the defect was insignificant and the vast majority of users would never even notice it” [4]. Clearly, this was not the case, and eventually the company offered to replace any flawed pieces with ones that did not have the defect. However, Intel had knowingly sold flawed processors, and continued to do so even after creating a version that had no such fault. They planned to release the fixed version only after selling their remaining stock of flawed chips [4]. I find this case to be very similar to my situation. Intel knowingly sold flawed pieces, just as my manager wishes to do. However, Intel had a public relations disaster after the Cost of Design Improvement 2 Ronen Orland The third case study I read is a fictional scenario. In it, an engineer for a small business designs a component for a company. Though it is designed to a satisfactory degree for the company, the component is not quite perfect and has a miniscule problem. The company orders 1000 of these components and production begin. After the first shipment, the engineer thinks of a solution to the small problem and learns that it can be implemented into the rest of the shipment, but would require slightly changing production and would add an extra three dollars cost per component [6]. At this point, the study asks the reader to consider the following possibilities: tell the company about the improvement and offer to share the expenses, pay for the expenses, or have them cover the expenses; or not tell the company about the improvement until after the entire order is shipped. I believe that not telling the client company of the improvement that would not cost much to implement now would be unethical; it would be withholding useful information from them for the improvement that they would almost certainly prefer to have over any defects. I would instead tell them about the improvement, as then they would decide whether they wish to implement it all. If they do, I would offer to either share the expense or cover it myself, as it would not be fair to suddenly propose a new idea that demands more money from them after an agreement has already been reached. After analyzing this study, I think that though it initially sounds applicable to my situation, I am instead working to avoid the one described in this case. Instead of knowingly producing faulty processors and then offering to fix them later, I wish to wait until the design is perfected. Also, the described scenario would not work for me as my company would have to completely replace the faulty processors, which would be much easier than attempting a repair on such a complex device. The second group of people I talk to are my family and friends who put together their own computers. I know that they are among the people who could be affected by the faulty processor as they are in the target audience for the product. I confide in them the state of the processor at the moment, and they are shocked that it is even being considered to be produced and sold given the safety risk regarding the short circuiting, as well as the possible damage to the rest of their computer. The defect could cause them to have to replace other components in their computers, which tend to run at rather high prices. The defect could also potentially damage their hard drives, which store all the data of the computer, meaning that they could also lose everything stored on their machines. They whole-heartedly agree with me and further convince me that I must speak again to my manager. SOLUTIONS AND CONCLUSION My coworkers and I decide that the first thing we should do is approach the department manager and talk to him about the plans for the processor. We would talk about exactly what the problems consumers could face when buying the product, namely: a product that works efficiently for only a fraction of the time that similar products typically do; a product that could potentially ruin other parts of their computer; and a possible safety hazard in the form of flames resulting from a short circuit from the product, as has been reported by individuals in similar situations [7]. We explain that it would be ethically wrong to purposefully sell such a hazardous product to our customers. We also talk about the repercussions of putting the current processor design into production. Eventually, users will find out the cause of what is short-circuiting their computers, and they will be furious. Not only would it cause a public-relations disaster, but people could also take legal action against the company. Lawsuits would only take more time and money to solve, and in the end the company would likely have to recall the product and reimburse everyone who was affected by the flawed product. We would conclude that all these situations could be avoided by simply giving more time to creating a better design. If the manager is still not convinced by our arguments, my team and I would go to the appropriate superior and present the same argument. If this also fails, then the least my team and I could do then is to not sign off or give our approval for the current design. A possible alternative is to begin immediately doing more research and work to create a safe and longer-lasting design for the processor. If this can be achieved quickly enough, then we could avoid putting the unsafe design out for production. However, this would be the least optimal solution as it would require rushing through the design process and potentially overlooking a different flaw or defect from lack of time to properly investigate. There is also the possibility that my team and I would not manage to ADDITIONAL SOURCES To aid me in my decision, I reach out to other, more personal sources. Coworkers The first people I approach are my coworkers. As they are working on this research and project alongside me, they will also be affected by any decision made regarding the processor. After briefing everyone on the results of the tests we ran on the processor design, they are in agreement that it is best to hold off on production until a better and safer design is created. We agree to work together to try and persuade whoever we must to do so. Family and Friends 3 Ronen Orland create a new design quickly enough and that production would begin anyway. Overall, the best solution that would be most ethical and professional would be to confront the manager and discuss the matter once more. I believe in face of all the facts and likely consequences, he would be persuaded and the situation would be avoided. REFERENCES [1] S. Wilber. (2015). “What is a short circuit?” PhysLink. (Website). http://www.physlink.com/Education/AskExperts/ae470.cfm [2] (2007). “Code of Ethics for Engineers.” National Society of Professional Engineers. (Online article). http://www.nspe.org/resources/ethics/code-ethics [3] (2015). “IEEE Code of Ethics.” IEEE. (Online article). http://www.ieee.org/about/corporate/governance/p7-8.html [4] C. Fleddermann. (2007, August). “Engineering Ethics cases for Electrical and Computer Engineering Students.” IEEE. (Online PDF.) http://www.ele.uri.edu/courses/ele400/f15/cases.pdf [5] M. Matteson, C. Metivier. (2015). “Case: The Ford Pinto.” Philosophia. (Online article). http://philosophia.uncg.edu/phi361-matteson/module-1-whydoes-business-need-ethics/case-the-ford-pinto/ [6] D. Zacker. (2006, August 23). “Cost of Design Improvement.” Online Ethics Center. (Online article). http://www.onlineethics.org/Resources/Cases/wonder.aspx [7] M. Dowler. (2009, February 11). “Beginners Guides: Most Common Ways to Kill a PC.” PCstats. (Online article). http://www.pcstats.com/articleview.cfm?articleid=1720&pa ge=7 ACKNOWLEDGMENTS I would like to acknowledge my friends for helping me think of a situation to fit my topic when I had trouble doing so. 4 Ronen Orland 5