This is a new browser window. When finished viewing, just close this window.

Universal Access to the Net: Requirements and Social Impact

Jeff Johnson

Invited talk at ACM 1997 Conference on Computer-Human Interaction (CHI'97)


This talk addresses the following questions: Where do we stand today with respect to achieving universal access to the Internet? What is required (particularly in the HCI realm) to achieve it? What are some of the consequences and side-effects -- positive and negative -- for society?


I'll start with a bit of status on where we are today. Then I'll list some requirements: what it will take to achieve universal access to the net. Finally, I'll talk about how universal access to the net might affect society. Ben Shneiderman (1990) called for social impact statements for technological developments. I agree and will do my part.

I can't be comprehensive because of time-constraints. I'll focus on issues that are relevant to Human-Computer Interaction: things those of us in the HCI-field can affect... and only some of those.


OK, first some statistics on where we are today in the U.S. My sources are given in the bibliography.

The statistics below show that Internet use is still pretty uncommon in comparison to other home-based technologies:

Another indicator of the status of universal net access is the number of civic networks. Civic networks are set up by and for local communities. They provide public access, local information & services: commercial, government, and civic. About 400 have been setup nationwide, and more are being established as I speak.

Here are several successful civic networks, starting with early ones and ending with recently established ones:

So we can see that Internet use in the U.S. is currently nowhere close to universal, but it is rising fast, so it seems that achieving universal access might be possible.


In contrast to the U.S., global universal access to the net looks remote indeed.

[Slide 1: third world village street]

As you've probably read because it's been circulating around the net, 3/4 of world's people have never used a telephone.

[Slide 2: broken-down bridge]

Many 3rd world countries have no road or sewage infrastructure to speak of.

[Slide 3: thatch-roof houses]

Information infrastructure is not next on their list. The 3rd world is spread out over our entire history... all the way back to B.C.

[Slide 4: Nepalese farmer plowing]

This is a farmer in Nepal, plowing his potato field with an ox and a biblical age plow made of a forked branch carved a bit to make it work better. His potato field is the size of your living room. With him in the foreground are his wife and malnourished child.

When someone predicts what life will be like in 2047, you should immediately ask: life for whom?

At ACM'97, moderator James Burke alluded to this when he said we are already living in the 3rd world's 21st century and can ask, for each part of the 3rd world, are we living in their 2007?

[Slide 5: woman and children selling limes]

Their 2017?

[Slide 6: coppersmith]

Their 2027?

[Slide 7: girls carrying wood]

or their 2047?

[Slide 8: monk & his elderly, soot-covered mother]

Of course, it's not as simple as that.

[Slide 9: man spinning wool wearing Nikes]

First, no inhabited place on earth is uniformly stuck in some past time period.

[Slide 10: Katmandu street w/oxen and motorcycles]

In the 3rd world as in the 1st, artifacts and practices from the past and the present are mixed together.

[Slide 11: adobe store in Peruvian Andes with Pepsi and Coke signs]

It's just a matter of which period dominates.

[Slide 12: kids tending babies, watching helicopter]

Second, Burke's observation is oversimplified because it assumes that more advanced technology means a better life. So-called "primitive" life can be OK if you have food, health, peace, and pleasant surroundings.

[Slide 13: native flutist in Andes]

A corollary is that bringing people into the modern age by putting them on the receiving end of high-tech weapons or technologies of oppression isn't really an advance for them.

Third, Burke's observation is oversimplified because it assumes that if some locale is 50 years behind us now, it will progress in 50 years to where we are now. But that of course is far from guaranteed.

Oversimplified though it may be, Burke's observation helps remind us that access to the net is nearer term for this little boy:

[Slide 14: American boy at computer]

than for this one:

[Slide 15: dirty-faced boy in Nepal]

I don't believe global universal access to the net is possible, by 2047 or any other time. Some rulers won't let their people participate. Some people, some entire cultures, won't want to participate. The best we can expect is universal access in the 1st world, and fragmented access for the rest of the world.

So I wish technology pundits wouldn't say "...everyone will do X..." when it won't really be everyone.


Now, some requirements.

I'll quickly list some requirements for achieving universal access that have little to do with human-computer interaction:

Now, some HCI-related requirements.

Security and Privacy

An important one is that security and privacy measures must be available. Most of this has nothing to do with Human-Computer Interaction, but there is an HCI component. Secure encryption and authentication is available today: you can use PGP for example. But it isn't yet user-friendly. For people to feel safe in using the net, security and privacy measures need to be invisible... built into applications in application-specific ways.


Another HCI-related requirement is that the net must provide value. As Tom Landauer and Steve Miller have argued, it must be not just usable, but also useful.

It should have been obvious, for example, that the Interactive TV trials of the early 90s (e.g., Interactive Network) would fail: They allowed subscribers to guess football plays, play Jeopardy against guest stars, etc. They provided nothing people wanted. But many people bet serious money on them... and lost their shirts.

What constitutes value? What do people want? I see four possibilities: information access, commerce, entertainment, and communications.

Information Access

Academics and educators stress information access.

The buzzwords here are: online encyclopedia, and world of information. Others claim that information access is something only academics and intellectuals are interested in, and that it will never be a big part of net use.

I say: If we teach our kids to use it, they will. If we make information access relevant to everyday life, people will use it. For examples of what I mean, see my net scenarios in the readings.

Inhibitors to widespread use of the net for information access are: a lack of useful information, copyright limitations, costs, weak search mechanisms, and slow downloading. Another inhibitor is usage monitoring: if people think they are being tracked as they use the net, they won't use it much.

Commerce and Entertainment

Business people stress commerce and entertainment as the primary uses of the net.

The buzzwords here are: online shopping, video on demand, games on demand, distributed games.

My view of this is: If you have a hammer, everything looks like a nail. Business people think of the net in these terms because it's what they know. They know how to charge for it, it's much like today's market (push-oriented), and it doesn't require a thick backchannel for interactivity.

The inhibitors here are: a lack of critical mass: sellers, buyers, goods (like a new flea market: nobody's there because nobody's there), and poor search and delivery mechanisms.


I believe what people really want from the net is new ways to communicate with each other.

Here, the buzzwords are: email, chat, newsgroups, b-boards, and muds.

I think the "killer app" everyone is looking for is right under our noses. I think the other uses of the net (information access, entertainment, and commerce) will follow interpersonal communication. Maybe they'll catch up eventually, but communication will be what hooks people in. But how do you make money providing that? (Ah, there's the rub.)

On the other hand, one might claim that person-person communications is entertainment. If you look at what's posted in newsgroups, email, and chat rooms, most of it is just people making conversation. (One might say: Smalltalk took over after all, just not in the form Adele Goldberg hoped for.) Maybe new successful forms of net-based entertainment will be new forms of person-person communications (see the article "Not a Highway, But a Place" in the readings).

Conventional Usability

An obvious HCI-related requirement for universal access is usability: using user-centered design methods, designing for all ages and physical abilities, etc. I needn't say much about that in this crowd. But there is a special usability requirement I want to emphasize.

Easy Setup

People have got to be able to bring it home, plug it in, turn it on, and it works. We have a long way to go on this one: setup difficulties are a huge obstacle. To get online, people have to master: cables, connectors, installation glitches, readme files, drivers, fonts, software conflicts. I'm going to harp on this for a few minutes because this is something we in the HCI community ought to be able to do something about.

First some anecdotes, then some real data.

A year ago, I signed up for an online service at home. From the day I first tried to install the software and log on, to the day I actually logged on took one month. A month of calling customer support every few days to try their next theory. I ended up taking my computer to their office and letting someone there find and fix the problem.

I recently bought Bill Gates' book, The Road Ahead. It comes with a CD-ROM that lets you read the book online, follow hypertext links, see videos of Bill Gates, etc. The instructions for installing the CD are very simple... appropriate for a book about how great it will be when we're all online. Unfortunately, they didn't work. When I tried to open the book, a dialog box appeared reading: "Cannot find VBRUN300.DLL." After trying for a few days to find the missing file, I called the customer support number. The fellow said: "Oh, that file is supposed to be installed from the CD. I guess it wasn't for some reason." For someone like me, this type of thing is an annoyance... a temporary stumbling block. But for people who have no technical training, such problems are insurmountable obstacles.

My brother Mark, a businessman and father with two kids, sometimes calls on me to help him with PC problems. He says: (quote) "Installing software is the worst, most time-consuming and frustrating part about owning a computer. It never goes as the instructions say it will." And he owns a Mac.

I recently volunteered at a booth in downtown SF to answer people's questions about the Internet. It was set up by NetAction, a local online advocacy organization I'm involved with. The booth had hundreds of visitors over several days. People who stopped by mainly wanted hand-holding in solving their problems getting online. They also wanted background understanding: What's a web page? What's email vs. b-board vs. chat?

Here's some real data, from the December 1996 Communications of the ACM.

In 1995, US West conducted a trial of online services in Winona, MN. They had set up a civic net centered around schools to connect schoolkids, teachers, and parents. They provided families with training on setting up and using the system.

Their results: "The biggest obstacle was setup ... Two months after training, many families had still failed to get set up ... Families rated setup tasks more difficult than Internet tasks." It took housecalls evenings and weekends by support people to get everyone online.

Also in 1995, Carnegie-Mellon University conducted a trial in Pittsburgh. They lent or sold (at reduced price) PCs and modems to 48 families. They supplied a simplified Internet access system: a custom family Web page providing access to email, Web, newsgroups, and MUDS.

Their results: Again, setup was a big problem. Obstacles included: bad phone lines, busy signals, poor user interfaces, buggy software, forgotten passwords, erased logon scripts, depressed shift-lock keys, you name it.

Here are some observations common to both studies:

Reed Hundt, Chair of the FCC said at the ACM'97 conference: "The number of people who tried the Internet and gave up is equal to the number of Internet users. No other popular technology has that drop-out rate."

We in the HCI field have got to do something about this! We often help design and test the applications, but are rarely involved in designing and testing the installation process. We should take the study results and Mr. Hundt's comment as a challenge and get to work on this. Achieving universal access depends on it.


Enough requirements. How might achieving universal access affect society? In many ways, but I have time to mention only a few.

Effect on Literacy

How will it affect literacy?

Some people foresee a resurgence of writing skills from increased use of email. I'm not so sure about that. So far we seem to be seeing how many people are willing to broadcast their quasi-literacy or carelessness company-wide or worldwide.

More importantly, advances in multi-media may mean changes in what it means to be literate.

A couple of years ago, a reporter posed a thought-experiment:

Question: Does Emily need to be able to read and write, or will other skills have become more important for communication?

I had two separate analyses of this, one concerned with the state of society and one concerned with the state of technology. Both were intentionally negative.

First, I projected forward from today's declining support for taxes, public schools, and other public programs, and posited that society in 2020 is heavily stratified.

Some people live well, with technology,

[Slide 16: American girl at computer]

but most live in medieval squalor, without it.

[Slide 17: Tibetan ragamuffin girl]

Life is unstable: crime is rampant, and there are frequent uprisings, riots, guerrilla attacks, and kidnappings. The rich and middle class live behind intense security, and endure frequent and unpredictable infrastructure outages. Sort of like Lima, Peru today.

So my answer to the reporter's question was: if Emily is well-off, she at least needs to be able to read and write for times when the net is down. If she's poor, she definitely needs to be able to read and write because she has no access to technology, but ironically, she probably can't because she's had no education.

For my second analysis, I projected today's technology forward, and suggested that Emily will have a new Sony multi-media story player, but:

[Slide 18: Girl w/player]

So my answer based on this analysis is: Yes: for entertainment, Emily will have to read the books her grandmother left her.

As I said, my projections were intentionally negative. I hope we can avoid them and I challenge you to try. But if we do, will that mean that reading and writing won't be needed?

I expect that as the future unfolds, we will avoid some aspects of my projections and not others, and that "literacy" may mean something other than what it means today. What will it mean? What do we want it to mean? Should it matter what we, rooted in the 20th century, want it to mean? I think it should.

We might ask the same question about numeracy? Will people need to be able to do arithmetic? Do we need to now?

Effect on Spamming, Junk E-mail

How will having many more people online affect the dataglut and junk mail many of us experience?

I worked at Xerox in the early 80s. Xerox was one of the first companies to use email extensively for internal communication. There were distribution lists corresponding to the organizational hierarchy. There was a problem of occasional spamming (although it wasn't called that then), such as employees sending messages to all Xerox (tens of thousands of employees worldwide) selling Amway products or offering free kittens.

A colleague expressed concern that when email spread widely outside of Xerox (which it now has) some people would send messages to the world. I argued then that the lack of a world list or other similar lists made that impossible.

Well, I was wrong: currently, large lists are being compiled and abused. It is possible to send a message to all America Online subscribers, and enough advertisers do it that a friend of mine just switched from AOL: he got tired of receiving five ads a day. Even just browsing the Web can land you on dozens of e-mail marketing lists.

Economics of Spam

Let's consider the economics of spam. For any mass mailing, the factors limiting the number of recipients are the costs of list-compilation and of duplicating and mailing. For postal mail, duplicating and mailing costs are significant. This limits the amount of postal junk mail you receive (even though you may not consider it limited).

With email, duplication and mailing are free, so spamming is more likely. The only limiting factor is the cost of collecting or otherwise obtaining mailing lists.

Sometime when I argue against collection of consumer data, my opponents respond that consumer privacy is the problem, not the solution, that is, compiling more info about consumers would reduce the amount of junk mail because it would let advertisers target only those who are truly interested.

This argument is valid for postal junk mail, for which duplicating and mailing costs are significant. Advertisers don't want to waste money mailing stuff to you if you're just going to toss it. But with email and newsgroups, advertisers save nothing by targeting messages narrowly. They might as well spam everyone they can just in case someone might be interested.

The bottom line here is that data abuse and spamming must be addressed, especially as more people come online, or the net won't be a pleasant place. How might we address it?

Combatting Spam

One approach is social: develop, teach, and promote cyber-responsibilities and cyberethics: netiquette. We shouldn't assume that everyone who comes online just naturally knows netiquette. This is especially important for young people.

At the same time, we can pursue technological anti-spam measures. We in the HCI field can help to develop better searching methods, removing the pressure to make the net more push-oriented. HCI professionals can also develop better email filtering mechanisms: 'cause we're gonna need 'em!

Others in the computer field can help by developing measures to thwart list compilation. Or perhaps Internet service providers could consider charging postage per recipient, thereby introducing the same factor that limits postal junk mail.

Finally, we can try legislative and regulatory solutions, such as limiting spamming by commercial advertisers, or limiting collection and sharing of consumer data.

Effect on governments

What effect might a largely online 1st world have on governments? Some say the net makes political borders irrelevant, and so makes governments and nations obsolete. For example:

Is undermining government good or bad? Those who believe that the rise of the net spells the end of the nation-state fall into two camps:

Given a choice between the two, I'd tend to agree with Schiller.

But it isn't just a matter of choosing between the two. There are people who are trying to use the net, or parts of it, to improve and support government. I'm guardedly optimistic about local civic networks -- if we can get enough of them built -- fostering democracy at the community and town level. I am less optimistic about the wider Internet's ability to enhance democracy at the state and national levels, but I think that we have to try; because if we don't, the future will look more like Schiller's nightmare than Barlow's dream.

Effect on culture

Predictions about the net's effect on cultures parallel those about government. Some say the net will contribute to globalization of culture. Which culture?

Will all net users have to learn English? Will they all be trained to be good consumers? Will all societies have to embrace western human-rights ideals?

Again, good or bad? It depends on whether you view globalization as a cultural steamroller or a spreading of enlightenment. One person's enlightenment is another's steamroller. See Barber's book McWorld vs. Jihad, which predicts either the triumph of corporations and western consumerism, or a fundamentalist and nationalist backlash against it.

The challenge for HCI professionals is to help us avoid both McWorld and Jihad by developing technologies for bridging cultures without homogenizing them.


Bill Gates wrote: "No one gets to vote on whether technology is going to change our lives."

My response is that he is both right and wrong. He's right that the choices will not appear on any ballot. However, he is also wrong because making a choice on a ballot is not the only way to vote.

People vote all the time, by their actions and words. My brother voted by choosing a Mac over a Windows machine. I've voted by writing OpEd articles, organizing conferences, and advising regulators and legislators. I'm voting right now by giving you my opinions. Citizens around the country voted by establishing civic networks in their communities. A dictator may vote by decreeing that the Internet won't be accessible to his people.

If there's any message I want you to take away from this, it's that the future isn't something out of our control that just happens. To a large extent, we construct it. Sometimes we can't see what we are constructing, but we are constructing it nonetheless.

Those of us in the computer field have a great deal of leverage -- much more than the average citizen -- in determining the future because we design tools, infrastructure, and processes that are central for citizens of the 1st world, and that affect even people in the 3rd. This gives us a responsibility to use our leverage wisely.

Please get out there and vote.