i have hw on philosophy

Holding Safety Paramount

William S. Larkin

I. Duty to the Public

A. Intrinsic relation to the public

1. Engineering essentially involves solving complex material problems, and this requires producing technology--machines, devices, systems, etc.

2. The ‘public’ can be defined as those who use or are directly affected by uses of technology.

3. So one of the constitutive roles of the engineer is a relationship to the public.

4. Being a good engineer thus requires playing that role well.

B. Codes of engineering ethics

1. All of the codes of engineering ethics identify safety as a paramount concern.

2. To play the role toward the public well is to design/operate/maintain safe technology

C. Distinctive duty

1. The duty to the public to hold safety paramount is assumed to be duty that all engineers have in virtue of their being engineers

2. And the duty is one that is thought to go beyond what law, market, pragmatic, or general moral considerations might require

3. So engineers have a distinctive duty to hold safety paramount, in the sense that it applies to all and only engineers as such

D. Illumination required

1. Engineers need to know how to apply the general rule to hold safety paramount in particular cases--this requires understanding the meaning of the rule (its spirit rather than just its letter).

2. Engineers need adequate professional motivation to abide by the rule to hold safety paramount--this requires appreciating the justification of the rule (how it follows from the nature of engineering).

II. Naive Views

A. Minimal Risk

1. Intuitively, the idea is that a technology is safe if the level of risk is minimal—in some sense it is at the lowest level possible.

2. Risk is function of both the magnitude and the probability of harm.

3. Risk can be minimal in either an absolute sense or in a relative sense.

B. “Absolutely” Safe

1. On this view, to hold safety paramount is to insure that there is at least no significant risk--i.e., insure that the probability of any serious harm is very low and that the more likely harms are relatively trivial.

2. But no technologies needed to solve complex material problems can be completely without significant risk.

3. So this is an unreasonable standard for holding safety paramount.

C. “Relatively” Safe

1. On this view, safety considerations are supposed to override all others--to hold safety paramount means to favor safety whenever it conflicts with some other value.

2. But other values are essential to solving complex material problems--like effectiveness, efficiency, reliability, etc.

3. So safety can be legitimately traded off for these other values, at least to a point.

4. Allowing safety to trump every other value would effectively prevent engineers from being able to solve many if not most complex material problems.

III. Standard Views

A. Acceptable Risk

1. Intuitively, a technology is safe to the extent that it is not risky (safety and risk are inversely proportional)

2. The point at which no more safety can be legitimately traded off for some other value is the point at which risk becomes unacceptable.

3. So it seems we could perhaps define holding safety paramount in terms of ensuring acceptable levels of risk

4. But “acceptable” is ambiguous:

a. Descriptive sense: something count as ‘acceptable’ just so long as people are able to accept it, they would be willing to accept it

b. Normative sense: something counts as ‘acceptable’ just in case it is worthy of being accepted, people should be willing to accept it

5. These two interpretations of acceptable provide two broad variations of the standard approach to defining safety in terms of acceptable risk.

B. Compensation Approach

1. The view

a. A technology is safe if and only if people should be willing to accept the level of risk associated with it

b. People should be willing to accept the level of risk associated with a product if it is adequately compensated for by utility ( a function of the probability and magnitude of benefits).

c. SO: A technology x is safe if and only if P(Bx) > F[P(Hx)]

2. PROS

a. Using the normative sense of acceptability--acknowledges that people are not the best at discerning, appreciating, and valuing probabilities and magnitudes of benefit and harm themselves

b. Fits with a utilitarian approach to ethics--safe technology would be that which promotes the common good in the form of the greatest happiness

3. CONS

a. In practice the approach will tend toward being overly paternalistic--a group of technological elites will be telling everyone else what technology they should and should not have

b. In theory it doesn’t acknowledge that people have a right to autonomy and to take on risks for themselves

c. There are worries about being able to predict and assign probabilities to the uses of technology and then being able to quantify harms and benefits using a common metric

d. There seem to be cases where we would say that a technology is not safe even though the utility does actually outweigh the risk (e.g., the Pinto, perhaps)

C. Consent Approach

1. The view

a. A technology is safe if and only if people would be willing to accept the level of risk associated with it

b. People would be willing to accept the level of risk associated with a product if they would consent to being exposed to that level of risk if they had all of the relevant information

c. So: A technology is safe so long as the level of risk is above the level at which people would begin to withhold their informed consent

2. PROS

a. Using the descriptive sense of acceptability--avoids the tendency toward being overly paternalistic

b. It acknowledges public’s right to decide for itself (autonomy) what risk to take on

3. CONS

a. Tends toward overly instrumentalist view of engineering--engineers are just tools of the public interest (as filtered through/determined by government or the market)

b. Doesn’t acknowledge the systematic difficulties people have of evaluating probabilities and assessing risks

c. Cannot realistically obtain everyone’s informed consent, and difficult to model and predict what people would be willing to give their informed consent to

d. Cases where it seems that a product is unsafe even though people would be willing to give their informed consent (e.g., the Pinto, perhaps)

D. General Problem for the Standard Approach

1. If engineers have a distinctive duty to adhere to the standard of holding safety paramount, then engineers as such should have the technical competence to apply the standard and the professional motivation to abide by the standard in particular cases.

2. On either the compensation or consent approach engineers as such do not have the technical competence to determine whether a particular technology is safe.

a. On the compensation approach, expertise in history, sociology, and economics is required to determine whether risks are adequately outweighed by utility.

b. On the consent approach, expertise in psychology, statistics, or market research is required to determine whether public would consent to some level of risk.

c. Engineers as such do not have technical competence in the social sciences.

3. On either the compensation or the consent approach, engineers as such do not have the professional motivation to hold safety paramount.

a. On the compensation approach, the motivation to insure that risk is adequately outweighed by utility is either to promote the common good (idealistic) or to promote the welfare of one’s company or client (cynical).

b. On the consent approach, the motivation to insure that the public would consent to risks is either to respect the public’s rights to autonomy and informed consent (idealistic) or to protect one’s company and client from litigation (cynical).

c. Neither general morality nor pragmatic self-interest is a professional motivation for engineers as such.

4. So, if we accept either approach to the standard view of defining safety in terms of acceptable risk, then we cannot say that engineers have a distinctive duty to hold safety paramount.

IV. Alternative View

A. Optimal Risk

1. Another intuitive approach to what it means to hold safety paramount is to maintain that a safe technology is one with an optimal level of risk—not so little risk that the proper function of the product is undermined, but not more risk than is in some sense unavoidable given the essential character of the product.

2. So a safe technology is one in which there are no alternative designs available with a significantly better risk profile.

3. “Alternative” designs are ones that solve the given complex material problems within the constraints of the project.

4. “Available” designs are ones that a competent engineer could reasonably be expected to come up with in the circumstances with due diligence.

5. ‘Better risk profile’ has fewer of those system features that are strongly correlated with risk.

B. Risk Profile of a technological system

1. Need a way of rating and ranking the safety of alternative designs without having to determine the probabilities and magnitudes of harm associated with the various possible uses of the product (or predicting how people will react or feel about the product).

2. So we need to identify something that correlates strongly with risk but that is discernible by looking at the technological system itself or at its immediate inputs and outputs.

3. There is a connection between risk and uncertainty—most unintended outputs of a system are going to turn out badly.

4. Unintended outputs are primarily a function of (a) unintended inputs and (b) unreliable connections between inputs and outputs.

5. Unintended inputs can be managed by means of “idiot proofing,” and unreliable connections can be managed by means of “fail-safing.”

6. But engineered safety devices tend to increase both tight coupling and non-linear connections; and according to normal accident theory these system features are strongly correlated with unintended outputs.

7. So a better risk profile would have to have a better balance or blend of engineered safety devices, loose connections, and linear connections.

C. Competent Engineering

1. Careful

a. define complex material problem and the functional profile of a system needed to solve that problem

b. identify all relevant variables, constraints, and values

2. Creative

a. wide and variable array of potential design solutions

b. elegant structure of parts, units, and sub-systems

3. Conservative

a. relies on tried and true solutions to sub-problems

b. doesn’t over-design (keep it simple)

c. is attentive to details

d. has a feel for materials and operations of the system