Father of intel pentium chip

Intel’s Pentium Chip Crisis: An Ethical Analysis

Cindy Williams

Abstract—In October 1994 a mathematics professor informed

the Intel Corporation that its Pentium chip had a flaw which

caused mathematical errors. Intel’s response to the professor and

its customers created a backlash of anger and a public relations

crisis. By analyzing Intel’s actions using the work of two relevant

ethical philosophies, this article shows that some of the company’s

errors in public relations were also ethical errors. However, it also

points out that Intel has made improvements which will help it

avoid future problems and which could set an ethical precedent

for the semiconductor industry.

Index Terms— Communication ethics, Intel, Kantian ethics,

Pentim, public relations, utilitarianism.

Question: How many Pentium designers does it take to screw

in a light bulb?

Answer: 1.999 042 740 17. That’s close enough for nontechnical

people [1, p. 11].

AWORST CASE scenario for any business is to have

its credibility fall so low that it becomes the subject

of jokes among its customers and the general public. This

nightmare came true for one large company which has literally

millions of customers: Intel Corporation, a leading computer

chip manufacturer. Intel’s chip designs are inside 80% of

personal computers [2]. Intel’s nightmare began when one of

its chips, the Pentium processor, was found to have a flaw in

the portion that does math computations. While the actual math

errors affected a small segment of Pentium users, including

scientists, engineers, bankers, and the like, Intel’s response

to its customers’ concerns about the chip affected many

others, the majority of whom are “nontechnical people.” These

customers expected a perfectly working chip; Intel insisted that

perfection was not necessary. The media got involved and

Intel finally relented, but not before jokes like the one quoted

above about the Pentium’s mathematical inaccuracy and the

company’s unresponsive attitude began to appear.

Most accounts of the chip crisis describe Intel’s actions from

a public relations perspective. Press articles frequently quote

crisis management and public relations experts in their analyses.

A public relations journal analyzed Intel’s PR decisions

alongside McDonald’s responses to that company’s hot coffee

crisis [3], and an article in a business journal described the

Pentium ordeal as a “public relations problem” [4, p. 152].

This account will analyze Intel’s actions from a different

perspective—an ethical perspective. Relatively few published

responses to Intel’s crisis dealt with the ethics implications.

Manuscript received March 1996; revised October 1996.

The author is with Boise State University, Boise, ID 83725 USA.

Publisher Item Identifier S 0361-1434(97)01826-2.

One commentator stated, “The thing Intel seems to have

missed is the ethical issues involved. Intel’s responsibilities

to its customers and its customers’ customers are real and

profound” [5, p. 26]. The company’s ethics were also alluded

to by an analyst who said that “ customers with a problem

ought to be accorded respect” [6, p. 67] and by a consumer

who complained on the Internet, “Intel has no right telling me

or anyone else what I need” [7, p. 14].

As these statements point out, Intel should have considered

its ethical obligations to its customers. But what, according

to ethical philosophy, were its obligations, and what effort did

Intel make to fulfill them? How well did Intel communicate its

efforts to its customers? By applying some ethical principles

relevant to this case, along with some corresponding public

relations principles, I will consider how some of Intel’s

problems could have been avoided and what changes the

company has made to prevent future problems. I will also look

at the way Intel communicated with customers, or sometimes

did not communicate with them, and how this affected their

ethical image during the crisis. Finally, I will consider some

of the broader ethical implications the crisis could have for

the computer industry.

BACKGROUND

One June day in 1994, Thomas R. Nicely, a mathematics

professor at Lynchburg College in Virginia, was working on a

project regarding prime numbers. Four of the five computers

he was using calculated the same answer. The answer from one

of the computers, however, was different, beginning with the

ninth digit to the right of the decimal. The computer with the

different answer was the only computer containing a Pentium

chip [8].

Professor Nicely worked and reworked his numbers for four

more months until he was sure that the problem was with the

Pentium chip. Finally, late in October 1994, he called Intel to

report his findings but was “brushed off” [8, p. A17]. Nicely

then communicated with other Pentium-user scientists over

the Internet, asking them to run some calculations to verify

his findings. They too found that their Pentiums created math

errors [8].

On November 7, 1994, the trade publication Electrical Engineering

Times described Nicely’s findings [9]. Intel admitted

sometime around this date [9] that its own engineers had also

discovered the Pentium’s math problems during the summer,

but the company had decided that since encountering the error

was so unlikely, it would not need to notify Pentium customers

[10]. Intel believed that “an average spreadsheet user could

encounter this subtle flaw of reduced precision once in every

27 000 years of use” [11, p. 18]. In Internet discussion groups,

0361–1434/97$10.00 ã 1997 IEEE

Authorized licensed use limited to: IE Xplore. Downloaded on May 13,2010 at 1:5:19 UTC from IE Xplore. Restrictions aply.

www..DoownnllooaaddPPaappeerr..iirr

14 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 40, NO. 1, MARCH 1997

Pentium customers complained among themselves about the

chip’s glitch and Intel’s failure to disclose it [12].

Then the mass media picked up on the story. CNN broadcast

an account of the Pentium chip’s flaw on November 22.

The New York Times and the Boston Globe contained articles

concerning the problem that same week [9]. The Thanksgiving

holiday found the CEO of Intel, Andrew Grove, and his senior

management team around a conference table, discussing the

escalating crisis [13].

Some at the conference table suggested a no-questionsasked

return policy. But Grove “disagreed because of the

consequences” [14, p. B1]. So, Intel decided upon a qualitative

return policy. If a customer wanted a replacement chip, he or

she would have to talk to people at Intel who would decide

whether the customer really needed one [10].

The company communicated its policy by first addressing

the disgruntled Internet customers [13]. Grove composed an

Internet message at home apologizing for “the anxiety created

among you” by the Pentium chip’s problem [15, p. B6]

and explaining the conditional return policy [13]. Grove was

unable to send the message from his home, however, so an

Intel scientist sent it for him, and since the message did

not originate from Grove’s Internet address, some recipients

questioned its authenticity [9]. Intel then had its press staff

communicate the policy to media reporters and had field

representatives contact its major business accounts [13].

Intel also set up an “800” telephone line so that customers

could call to voice concerns and/or request a new chip. Within

about two weeks of the Internet apology, calls to the hotline

had declined. But then a second wave in the crisis hit. On

Monday, December 12, IBM announced that it was halting

shipments of its computers containing Pentium chips [9]. IBM

had run some tests of its own and discovered that typical

spreadsheet users might encounter a division error every

24 days, rather than every 27 000 years as Intel predicted

[16]. Grove was astonished at IBM’s decision because the

company had given him no advance warning. Worried Pentium

customers began overloading the long-distance carrier lines

with calls to the hotline [9].

The following weekend, Grove read a commentary in the

San Francisco Examiner written by Nicely in which the

professor said he would have to inform his students that

Pentium machines were less than perfect. At this point, the

magnitude of the crisis hit Grove. The NewYork Times quoted

him: “I didn’t know the scope of the problem. I didn’t know

until the end what the real objection was” [9, p. D6].

On Monday, December 19, 1994, the crisis team at Intel

decided, after an all-day meeting, to rescind its conditional

replacement policy and offer replacement chips to anyone

requesting one. Grove stated: “Finally we decided, ‘This is

the right thing to do, both morally and ethically”’ [9, p. D6].

On Wednesday, December 21, 1994, Intel ran a full-page

ad in the Wall Street Journal and several other national

publications. It read as follows:

To owners of PentiumTM processor-based computers and

the PC community:

We at Intel wish to sincerely apologize for our handling

of the recently publicized Pentium processor flaw.

The Intel Insideâ symbol means that your computer

has a microprocessor second to none in quality and

performance. Thousands of Intel employees work very

hard to ensure that this is true. But no microprocessor

is ever perfect.

What Intel continues to believe is technically an extremely

minor problem has taken on a life of its own.

Although Intel firmly stands behind the quality of the

current version of the Pentium processor, we recognize

that many users have concerns.

We want to resolve these concerns.

Intel will exchange the current version of the Pentium

processor for an updated version, in which this floatingpoint

divide flaw is corrected, for any owner who

requests it, free of charge anytime during the life of

their computer. Just call 1-800-628-8686.

The ad was signed by Andrew S. Grove, Craig R. Barrett,

and Gordon E. Moore, President, Vice President, and

Chairman of the Board, respectively, of Intel [17, p. A7].

On the same date that the ad ran, Grove expressed his

uneasiness about the precedent that the new policy would set.

He said that it had “big implications for the company and the

industry” and that “if we live by an uncompromising standard

that demands perfection, it will be bad for everybody” [14,

p. B7]. However, in order to re-establish its credibility and

image, the company proactively supported the new policy.

In January 1995, Intel began operating hundreds of replacement

centers in conjunction with corporations who already

provided technical services. It also began collaborating with

computer vendors such as IBM and Compaq to replace chips

in their customers’ machines. To communicate directly with

the public, Intel established many more toll-free hotlines for

customers to use. And Intel assigned employees to work in

retail stores during the weeks before Christmas [18].

Finally, Intel announced that any flaws it found in the future

it would document, and Intel would inform others of how to

obtain the information on its home page on the Internet World

Wide Web. Also, Intel said that it would put identification on

future chips so that if a flaw were discovered, companies and

users could more easily identify affected chips [19].

During the last quarter of 1994, Intel’s profits were affected

by its having to set aside $475 million for chip replacements;

the company reported a profit of $372 million on sales of $3.23

billion. During the first quarter of 1995, however, it reported

profits of $889 million on sales of $3.55 billion. By April of

1995, fewer than 10% of the people who owned Pentium-based

computers had asked for a chip replacement [20].

RATIONALE FOR ETHICAL ANALYSIS

Andrew Grove’s statement that the unconditional replacement

policy was “the right thing to do, both morally and

ethically” implies that Intel wanted to act ethically. It also

implies that its earlier policy was not the “right thing to

Authorized licensed use limited to: IE Xplore. Downloaded on May 13,2010 at 1:5:19 UTC from IE Xplore. Restrictions aply.

www..DoownnllooaaddPPaappeerr..iirr

WILLIAMS: INTEL’S PENTIUM CHIP CRISIS: AN ETHICAL ANALYSIS 15

do” and that the new replacement policy would fix that

lapse. Intel’s subsequent actions reflected their recognition

that ethical practices and policies promote the long-term, best

interests of a business.

The role that ethics plays in business success is recognized

by many prominent companies. The Business Roundtable

organization published an ethics report using information it

had gathered from 100 member companies. In the summary,

the report states that “ there is no conflict between ethical

practices and acceptable profits. Indeed, the first is a

necessary precondition for the second” [21, p. 9]. The same

report detailed the ethical philosophies and practices of 10

large companies. The following statements from some of the

detailed reports explain why the companies emphasize ethics

and the advantages of doing so:

ethical aspiration and consumer acceptance are two

ideas of one coin rather than in conflict.—General Mills

[22, p. 42]

Most employees interviewed believed that there is ultimately

a positive correlation between ethics and social

responsibility on the one hand and profitability on the

other.—Xerox [23, p. 138]

During the two nightmare incidents of TYLENOL poisonings

and in the subsequent press reports of the

product’s stunning recovery, the J & J Credo was

frequently referred to as the contributing factor in the

Company’s exemplary behavior.—Johnson & Johnson

[24, p. 78]

The continuing success of these companies attests to the fact

that “business ethics” is not an oxymoron. As an executive of

Levi Strauss & Co. said in a recent speech: “Doing the right

thing from day one helps avoid future setbacks and regrets.

Addressing ethical dilemmas when they arise may save your

business from serious financial or reputational harm” [25, p.

507].

Since adopting ethical policies and practices is in the best

interest of a company, a company must be able to determine

whether its actions are, or would be, ethical. It must be able to

measure its behavior against accepted ethical standards. Ethical

philosophers have long described standards of acceptable

ethical behavior. Cavanaugh, Moberg, and Velasquez divide

systems of ethical philosophy into three main categories:

utilitarian theories (which evaluate behavior in terms

of its social consequences), theories of rights (which

emphasize the entitlements of individuals), and theories

of justice (which focus on the distributional effects of

actions or policies) [26, p. 365].

During the early stages of the Pentium chip crisis, Intel

appears to have based its actions on ethical standards derived

from a partial, sometimes flawed, application of utilitarian theory.

The crisis worsened when the company violated some of

the ethical principles derived from rights theories. Fortunately,

Intel took several steps to end the crisis that exemplified a

much better application of both utilitarian and rights principles.

The remainder of this article focuses on the ethics of the

major decision-maker in the Pentium chip crisis, Intel Corporation.

Obviously the company was not the only participant

in the crisis, however, and one could question the ethics

demonstrated by the actions of some of the other players.

For instance, some have suggested that Internet surfers and

the media unfairly exaggerated the Pentium chip’s problems

[27]. Some distributors expressed support for Intel regarding

its conditional return policy [28]. The crisis was a complicated

mix of ethical and unethical actions by a variety of people, but

Intel was always the central character.

ETHICAL ANALYSIS

The following discussions of utilitarianism and Kantian

rights explain the relevance that each of these ethical philosophies

has to the problems Intel faced during its Pentium chip

crisis.

Utilitarianism

John Stuart Mill, one of the most well-known utilitarian

philosophers, wrote: “The creed which accepts as the foundation

of morals, Utility, or the Greatest Happiness Principle,

holds that actions are right in proportion as they tend to

promote happiness, wrong as they tend to produce the reverse

of happiness” [29, p. 157]. Utilitarianism is commonly

summarized as being the promotion of “the greatest good

for the greatest number” [30, p. 62]. Promoting the greatest

good, according to Mill, “ includes not solely the pursuit

of happiness, but the prevention or mitigation of unhappiness”

[29, p. 164].

Utility is not a simple standard. Knowing how to promote

happiness for a group of people, such as customers, can be

difficult. But a business must keep its customers happy to

succeed. Philip Crosby, an expert in quality management and

author of Quality is Free puts it this way: “You have to

understand what they need and learn how to give it to them

It’s a full-time job figuring out what your customer wants”

[31, p. 14]. Pentium users’ anger at Intel’s actions indicates

that the company may not have understood its customers well

enough to figure out what they wanted and expected, and this

lack of understanding created an ethical problem. If customers

expect a certain level of quality and are disappointed in the

final product, their happiness is diminished. Moreover, if the

company promotes the product in a way that creates that

quality expectation, the company acts unfairly by delivering a

lesser quality product. Mill says: “ it is confessedly unjust

to break faith with anyone: to violate an engagement, either

express or implied, or disappoint expectations raised by our

own conduct, at least if we have raised those expectations

knowingly and voluntarily” [29, p. 204].

Intel may not have realized that its customers expected a

chip that worked perfectly. But its “Intel Inside” advertising

implies that you have the best when you have Intel inside your

machine. A flawed chip is less than the best. As one commentator

stated, “Even technological neophytes understand that a

computer chip is supposed to get math questions right” [32, p.

16]. Intel needed to realize the expectations it had raised and

live up to its reputation. As a San Francisco public relations

executive stated, “They have a good, solid reputation that they

Authorized licensed use limited to: IE Xplore. Downloaded on May 13,2010 at 1:5:19 UTC from IE Xplore. Restrictions aply.

www..DoownnllooaaddPPaappeerr..iirr

16 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 40, NO. 1, MARCH 1997

have spent years and millions of dollars to develop. That’s the

equity they have to protect” [16, p. D3].

Intel probably also did not realize the potential for harm

that the Pentium’s flaw held. According to the ethics of

utility, Intel was obligated to try to prevent harming others,

and mathematical errors might have caused harm to some

of its customers. The inaccuracy could have directly affected

such customers as scientists, bankers, and technologists [33].

However, it would have indirectly harmed others if it had

caused people to reach wrong conclusions during, for example,

drug testing or securities valuations [34]. If Intel had realized

the danger, it surely would not have kept the flaw a secret.

Besides understanding its customers in order to make ethical

decisions a company must also weigh the pros and cons

of alternative decisions and choose the one that will create

the greatest good. A business must compare the happiness

to be derived from one decision against the happiness to be

derived from other decisions. All alternative decisions need to

be evaluated this way [30]. Intel may have overlooked some

important factors as it weighed its alternatives.

One of the primary ways a business measures “happiness”

is through monetary costs. A business measures how much

happiness customers derive from its product by how much

they are willing to pay for it compared to how much they are

willing to pay for other products [35]. The company must be

able to offer the product at that price and still pay its suppliers

and employees. If it can, this is good for the customers, who

have what they want at a price they want; for the company,

which stays in business; and for its suppliers and employees,

who derive a paycheck. The company creates the greatest

overall good at the point of profit maximization. At this point,

it is producing the greatest quantity of product that its limited

resources allow; it is working most efficiently [36]. Efficiency

and profit maximization are based on the utilitarian concept of

creating the greatest good for the greatest number [35].

As the discussion above shows, a company must consider

monetary costs when evaluating its alternatives in a situation.

A change in the company’s resources can change how much

it can produce, how much it can pay its employees, and

sometimes its ability to stay in business. These things affect

people’s lives and happiness—the key elements of utility.

However, as we shall see, a company must consider other

costs besides just monetary ones as it looks at its alternatives.

One way that companies frequently evaluate alternative

decisions is by doing a cost/benefit analysis. They evaluate

the costs of a course of action to see if the costs outweigh

the benefits to be derived from that action. When Intel first

decided to keep the chip flaw a secret and then decided against

a full replacement policy “because of the consequences,” it

apparently compared the costs of replacing thousands of chips

with the benefit of solving such an “extremely minor problem”

by full replacement, and it concluded that the costs outweighed

the benefits. In terms of monetary costs, the decision may have

been right. A $475 million cost is certainly significant. And

Intel was convinced that it would benefit only a few customers.

Indeed, in the end it seemed to benefit only 10% of them.

When an action that will cost a great deal is for the good

of only a few, it may seem like the wrong action to take.

Tom Sorell makes this argument in his article, “The Customer

is Not Always Right.” Sorell contends that a business is

justified in asking: “Is deference to the customer likely to

cause business failure or significant loss of profitability while

preventing at most minor harm to the customer?” [37, p.

917]. If the answer is “yes,” the company may be forgiven

for declining to meet its customers’ demands. He cites as

a case in point the 1988 salmonella health scare in Britain

that devastated many egg producers. The media reported a

government agency’s suggestion that eggs posed a health risk,

but it turns out that the evidence was shaky. Sorell contends

that egg producers, had they had the chance, would have been

under no moral obligation to issue warnings or otherwise incur

costs in preventing a problem that might never occur [37].

In Intel’s case, the egg producers’ case, and other similar

cases, however, the question is really not “Will this cost us?”

but “How much will this cost us and in what ways?” For

no matter what decisions a business makes, and no matter

how unfairly it may be forced to make them, it incurs costs

in either monetary or nonmonetary ways. And sometimes the

nonmonetary costs can be very high.

Two nonmonetary costs that Intel needed to consider were

external failure costs and opportunity costs. External failure

costs are those that a company incurs when a product fails

once it reaches the customer. They include costs that can be

measured, such as returns, recalls, complaint handling, and

warranty servicing, as well as costs that cannot be measured,

such as a loss in future market share because of customer

dissatisfaction [38]. These costs may be very high for a company

because in competitive markets (chip markets certainly

qualify), a company can quickly lose market share if its

products do not conform to its customers’ quality expectations

[39]. So even though Intel was the only game in town for

Pentium chips, as soon as an alternative became available,

it would be in danger of losing its customers. IBM was one

customer that was apparently willing to wait for an alternative.

Intel was fortunate that it changed its policy before one came

along.

Another cost that is hard to measure but that a business

must consider in any decision is opportunity cost. When a

company expends resources to take advantage of one opportunity,

the same resources are then unavailable to pursue another

opportunity. So the opportunity costs of a decision are all the

other opportunities that the company foregoes in making that

decision [36]. In terms of external costs, whatever a company

spends to fix the problems caused by a substandard product

is then unavailable to spend on new product development,

advertising, other customer services, or anything else. The time

and money that Intel spent to conduct emergency meetings,

communicate over the Internet, operate the hotlines, put extra

staff in stores, respond to the media, and finally replace the

product could have been used for many other more productive

purposes. When Intel chose to distribute the flawed chips, to

keep the flaw a secret, and to deny its customers a replacement,

Intel missed three chances to lower its opportunity costs.

If Intel considered the external failure and opportunity costs

of its original policy decision, it underestimated them. One

cannot really fault Intel, however, if it tried to consider these

Authorized licensed use limited to: IE Xplore. Downloaded on May 13,2010 at 1:5:19 UTC from IE Xplore. Restrictions aply.

www..DoownnllooaaddPPaappeerr..iirr

WILLIAMS: INTEL’S PENTIUM CHIP CRISIS: AN ETHICAL ANALYSIS 17

costs but just did not have enough information to realize them.

Grove’s admission that he “didn’t know until the end what the

real objection was” indicates that lack of information probably

kept Intel from effectively applying utilitarian principles. If

this is the case, then while Intel initially made a wrong

decision, it did not have unethical motives [36].

The principles of rights theories require less information

to apply. These theories are based upon our duties to one

another. Companies frequently apply both utilitarian and rights

principles to come to ethical decisions, because the theories

complement each other by filling in considerations that the

other seems to miss [35]. Intel eventually recognized its

customers’ rights, and this recognition plus some further

utilitarian considerations helped the company solve its crisis.

Kantian Ethics: The Customers’ Rights

Grove’s statement of “the right thing” suggests that some

concept of goodness—some standard—may exist which is

not dependent upon a “greatest good” outcome. Perhaps it

was Intel’s duty to regard issues other than those of costs

vs. benefits. If the company had a duty to act some way in

response to the Pentium problem, its customers had a right to

expect that they would be the recipients of this action.

Ethical standards which are based upon our duties toward

one another may be explained by the philosophy of Immanuel

Kant. Kantian ethics are based upon two different variations of

Kant’s famous Categorical Imperative. He writes, “Therefore

there is only one categorical imperative, namely this: Act only

on a maxim by which you can will that it, at the same time,

should become a general law” [40, p. 170]. Within this principle

are the concepts of reversibility and universalizability.

Reversibility requires that behavior determined to be morally

right be moral whether it is directed toward yourself or toward

someone else. It “requires that you not make an exception of

yourself” [35, p. 19]. Stealing is easily recognized as immoral

with this test. Universalizability is similar but expanded. An

action is moral only if you and I and everyone could do it

and still have a desirable society. Lying is discouraged by this

test. If everyone lied to everyone, no one could ever know

who was telling the truth, and this would create an undesirable

society [35]. According to these principles, Intel needed to ask:

If we decide to disregard our customers’ concerns, would it

be all right for them to disregard ours? and, If we choose

to withhold information regarding product quality from our

customers, would it be all right for everyone (such as our

suppliers) to do the same to us? [30].

A company needs little information in order to ask these

questions and apply Kant’s principle. In fact, the principle is

almost intuitive. Kant says, “The common reason of men in

their practical judgments agrees perfectly with this and always

has in view the principle suggested here” [40, p. 150]. And

later: “Therefore, I do not need any sharp acumen to discern

what I have to do in order that my will may be morally

good. [As I am] inexperienced in the course of the world

and incapable of being prepared for all its contingencies, I

can only ask myself: ‘Can you will that your maxim should

also be a general law?”’ [40, p. 151]. Even without “being

prepared for all contingencies,” Intel could have thought about

fairness, which would have required that the company treat its

customers only as it would want or expect to be treated by

others. This was its duty; customers had a right to be given

this consideration.

Kant varies the categorical imperative to include the concept

of respect for the “will of a rationale being” [40, p. 175]

by writing: “Accordingly, the practical imperative will be as

follows: Act so as to treat man, in your own person as well

as in that of anyone else, always as an end, never merely as a

means” [40, p. 178]. In other words, people should be treated

as people, not as things or merely as the means to achieve

things. Velasquez gives a good explanation:

For Kant this means two things: 1) respecting each person’s

freedom by treating people only as they have freely

consented to be treated beforehand and 2) developing

each person’s capacity to freely choose for him or herself

the aims he or she will pursue. On the other hand, to

treat a person only as a means is to use the person

only as an instrument for advancing my own interests

and involves neither respect for, nor development of,

the person’s capacity to choose freely [30, p. 81].

As with utility, applying the principles of treating customers

fairly and valuing them as people apart from being the means

to an economic end is not easy. It can be expensive for a

company. For example, most of us are familiar with Johnson

& Johnson’s TYLENOL crisis. But that was not the only time

the company incurred a high cost to put its customers first.

Before the harmful effects of sun rays were widely known, J

& J began an ad campaign to market baby oil as suntan oil.

The ads were very successful, but when management at J &

J learned of the possible health hazard, they discontinued the

campaign, which resulted in a $5 million drop in sales [24].

But it did not destroy the company. J & J shifted its resources

into another opportunity and kept going [24]. The company

met both the Kantian imperative to treat its customers as ends

and the utilitarian standard of mitigating unhappiness.

Intel too needed to apply Kantian principles to deal fairly

with customers regarding the Pentium chip. For example, Intel

erred when it acted paternalistically toward its customers.

Intel’s policy of deciding who needed a new chip and who

did not demonstrated a “father knows best” attitude [28, p.

18]. Andrew Grove admitted that, “we got caught between

our mindset, which is a fact-based, analysis-based engineer’s

mindset, and customers’ mindset, which is not so much

emotional but accustomed to making their own choice” [14,

p. B7].

This angered consumers because they felt that they were not

getting what they had bargained for. The Wall Street Journal

quoted an Internet message from a customer in North Carolina:

“I’m mad as hell at the suggestion that Intel would ask me for

proof of what I’m using the chip for before offering to replace

it. I paid out my hard-earned dollars for a chip and I want

it to operate 100 percent or I want my money back” [12, p.

B4]. An Internet user quoted in Computerworld said, “Intel

should replace anyone’s chip that wants a new chip. The folks

paid for an IEEE floating point unit, and the current Pentium

does not meet that specification. Intel has no right telling me

Authorized licensed use limited to: IE Xplore. Downloaded on May 13,2010 at 1:5:19 UTC from IE Xplore. Restrictions aply.

www..DoownnllooaaddPPaappeerr..iirr

18 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 40, NO. 1, MARCH 1997

or anyone else what I need. It’s my money, not theirs. I buy

what I want” [7, p. 14].

Intel’s customers had a right to have what they paid for.

In order to obtain a Pentium chip, the customer had to meet

a price expectation. Likewise, in order to obtain a customer’s

money, Intel was assumed to meet a quality expectation. This

coincides with the previous utilitarian discussion about living

up to expectations that a company creates. When Intel failed

to do this, it failed to live up to the bargain. It did not respect

the right the consumer had to a chip that would “operate 100

percent.” Kant wrote, “Here there is no question of inclination,

only of the rights of others. It is not their needs that count in

this connexion, but their rights ” [41, p. 193]. Intel was

wrong to base chip replacement on its own definition of need.

It indeed owed a new chip to any customer that wanted one.

Intel corrected its error in implementing a by-request replacement

policy. Also, it decided to disclose future flaws

and let the consumer make his/her own decision regarding

quality. These policies were what Intel’s customers wanted all

along [4]. As evidenced by how few people actually requested

replacement chips, the real desire of its customers was to

“make their own choice.”

Intel needed to apply Kantian principles to its communication

with customers during the crisis as well. If we were

to review all of Intel’s communication errors, we would find

that several caused public relations problems. For example, in

response to Grove’s “electronic faux pas” [42, p. 1] of posting

his Internet apology from a different address, an article in

Advertising Age’s Business Marketing sports the embarrassing

headline “Intel wipes out surfing the ‘Net’” [42, p. 1]. If we

narrow our focus to the communication errors that caused

ethical problems, we can see instances where Intel could have

avoided the problems by applying the principles of respecting

the customer’s rights to be treated fairly and to make informed

choices.

The first communication error Intel made was deciding at

the outset to withhold information about the Pentium chip’s

problem. This decision was unethical based on the notions

of reversibility and universalizability. As well, withholding

the information denied its customers free choice. InfoWorld

states that the company usually “meticulously” documented

chip errors on errata sheets but chose not to with the Pentium’s

flaw [43, p. 18]. It should have followed its usual procedure

and documented the error, letting its customers decide how

significant the error would be to their applications. By coming

around to a full disclosure policy, Intel has taken action to

avoid this error in the future.

Another communication error involved Intel’s use of technical

and statistical arguments. Intel based many of its early

crisis management decisions on the statistical probability that

the Pentium’s flaw would be encountered only once every

27 000 years. Even as Intel apologized in national newspapers,

it maintained that the Pentium chip’s problem was “extremely

minor” and that its processors were “second to none.” It had

faith in its technology.

The company tried to use its technical expertise to deal

with the crisis. When Intel first established its “800” phone

line, it staffed the line with technologists to explain things

to customers and to determine who would qualify for a

replacement chip [13]. This created both public relations and

ethical problems as explained by Ian K. Mitroff, director of

the USC Center for Crisis Management: “This assumes that

everyone is a scientist or a statistician; numbers do not treat

people as humans and therefore alienate them even further”

[44, p. B7]. As we saw from Kant’s practical imperative, it is

important ethically that people be treated as humans. Barraging

people with numbers does not promote this principle. It is also

poor public relations.

INDUSTRY IMPLICATIONS

As a whole, the computer industry does not have a very

good record for delivering perfectly working products to its

customers. T. R. Reid, a Washington Post columnist, originally

tried to defend Intel’s conditional replacement policy, but later

understood his readers’ frustrations after being inundated with

letters from them. He explains the industry problem well: “The

underlying problem is that the PC business has developed

a tolerance for subpar products Flawed products are so

common that the computer magazines run regular columns

listing recently discovered bugs in best selling products [32,

p. 16]. Note that even in its apology letter, Intel said that “no

microprocessor is ever perfect.” And the man who blew the

whistle on the Pentium chip, Professor Nicely, has said, “All

chips have bugs. The chips we’re using now are some of the

most complicated we’ve every produced and I think they’ve

reached a level of complexity such that you can’t completely

debug them. You not only can’t eliminate all the errors, I

don’t even think in the lifetime of a chip it would be possible

to catalogue all the errors” [8, p. A16].

Will Intel’s experience change all of this? No, because insofar

as Nicely is right that it is impossible to eliminate all bugs,

we as consumers will still encounter them. However, Andrew

Grove’s prediction of “big implications for the company and

the industry” could still come true. Industry sources have said

that this is likely in the following ways:

• Intel’s full disclosure policy will cause others in the

industry, such as those who manufacture software and

computer-related items, to reveal bugs in their products

too [45]. While a product may not be error-free, consumers

will at least be able to decide for themselves which

errors are important to them. In May 1995, just following

Intel’s crisis, Hewlett-Packard announced that it had

found a flaw in some of the computer equipment it had

recently sold, and the company made this announcement

before it had received any customer complaints about the

problem [46]. This type of disclosure exemplifies what

could become a new industry standard.

• Companies in the industry will take the customers’ views

more seriously. This crisis has shown how indignant

customers can create, as Ed Foster, editor of InfoWorld

calls it, a “very big hammer, one that no company is too

powerful to ignore” [27, p. 52]. Consumers have proven

that they can force themselves to be heard by companies

who do not listen to their concerns. As one analyst stated,

Authorized licensed use limited to: IE Xplore. Downloaded on May 13,2010 at 1:5:19 UTC from IE Xplore. Restrictions aply.

www..DoownnllooaaddPPaappeerr..iirr

WILLIAMS: INTEL’S PENTIUM CHIP CRISIS: AN ETHICAL ANALYSIS 19

“There’s much to be learned from the Pentium problem,

but most important of all is that the consumer is boss,

like it or not” [28, p. 18].

CONCLUSION

Acting ethically is in Intel’s, or any company’s best interests.

Intel made some ethical errors by not respecting the rights

of its customers. The company also made utilitarian errors,

some of which were probably caused by lack of information.

The most basic information can be obtained, however, by

asking questions based on Kantian principles, such as, “If I

act this way toward my customers, would it be all right if

they (and others) act the same way toward me?” The answer

would have prompted Intel to respect its customers’ rights and

to implement its final policy at the start.

To its credit, Intel finally came around to respecting its

customers’ rights, which correspondingly promoted their happiness.

When it decided to give a new chip to anyone who

requested one, to disclose all future chip flaws, and to communicate

its concern and solutions to the public, Intel reflected

the ethical principles of fairness and the pursuit of the “greatest

good.” And if the Pentium chip crisis serves to promote these

principles industry wide, this will increase happiness for all of

us, “nontechnical people” and “technical people” alike.

REFERENCES

[1] “Take my Pentium chip. Please,” New York Times, sec. 3, p. 11, Dec.

18, 1994.

[2] S. K. Yoder, “Technology: The Pentium proposition: To buy or not to

buy,” Wall Street J., p. B1, Dec. 14, 1994.

[3] A. Gonzalez-Herrero and C. B. Pratt, “How to manage a crisis before—

or whenever—it hits,” Public Relations Quart., vol. 40, no. 1, pp.

25–29, Spring 1995.

[4] N. R. Augustine, “Managing the crisis you tried to prevent,” Harvard

Bus. Rev., vol. 73, iss. 6, pp. 147–158, Nov.–Dec. 1995.

[5] M. Gibbs, “Intel Inside: An unknown quality,” Network World, p. 26,

Dec. 5, 1994.

[6] H. Beh, “Lessons in the price of arrogance,” Asian Bus., p. 67, Mar.

1995.

[7] J. Vijayan, “Intel miscalculates Pentium user backlash,” Computerworld,

pp. 1+, Dec. 5, 1994.

[8] P. Baker, “Hello, Mr. Chips: Va. teacher who found Intel’s flaw,”

Washington Post, pp. A1+, Dec. 16, 1994.

[9] J. Markoff, “Intel’s crash course on consumers,” New York Times, pp.

D1+, Dec. 21, 1994.

[10] , “The chip on Intel’s shoulder,” New York Times, sec. 4, p. 6,

Dec. 18, 1994.

[11] B. Crothers, “Pentium woes continue: Faulty FPU flubs math in certain

equations,” InfoWorld, pp. 1+, Nov. 28, 1994.

[12] D. Clark, “Some scientists are angry over flaw in Pentium chip, and

Intel’s response,” Wall Street J., p. B4, Nov. 25, 1994.

[13] E. Corcoran, “The glitch that stole Intel Corp.’s Christmas,” Washington

Post, pp. A1+, Dec. 19, 1994

[14] J. Carlton and S. K. Yoder, “Humble pie: Intel to replace its Pentium

chips,” Wall Street J., p. B1+, Dec. 21, 1994.

[15] D. Clark, “Intel’s Grove airs apology for Pentium over the Internet,”

Wall Street J., p. B6, Nov. 29, 1994.

[16] L. Helm, “Intel’s handling of the Pentium defect chips at its image,”

Los Angeles Times, pp. D1+, Dec. 13, 1994.

[17] Intel Corp. Advertisement, Wall Street J., p. A7, Dec. 21, 1994.

[18] K. Bertrand, “Intel starts to rebuild,” Advertising Age’s Business Marketing,

pp. 1+, Feb. 1995.

[19] B. Crothers, “Intel learns from Pentium problem, changes policy on chip

flaws,” InfoWorld, p. 37, Feb. 6, 1995.

[20] T. Quinlan, “Intel recovers from Pentium woes with record profits,”

InfoWorld, p. 12, Apr. 24, 1995.