The Internet’s
Future Is More Fragile Than Ever, Says One Of Its Inventors
Vint Cerf, the co-creator of tech that makes the internet
work, worries about hacking, fake news, autonomous software, and perishable
digital history.
I’m squeezed into a side corridor at the
elegant Westin St. Francis hotel in San Francisco during the Association for
Computing Machinery (ACM) annual conference in June. Sitting across from me is
Vinton Cerf, whose moniker, “Father of the internet,” might as well be part of
his legal name. In his signature style, Cerf wears a three-piece suit with
pocket handkerchief, his white beard and hair closely cropped. Despite his raft
of awards (including the ACM’s Turing Prize in 2004 and the Presidential Medal Of Freedom
in 2005), and position as chief internet evangelist
at Google, Cerf is disarming, modest, and often funny.
Cerf doesn’t demur the title Father of the
Internet, but he emphasizes that he co-parented with others–especially Robert
Kahn, with whom he has shared major awards. In the mid-’70s, Cerf and Kahn
crafted TCP/IP, which became the internet addressing system that ensures a data
packet created anywhere in the world gets to a computer, smartphone, robot,
smart car, drone, or connected thermostat anywhere else in the world—although
many of the gadgets now on the net weren’t even glimmers in a futurist’s eye at
the time. “Bob and I didn’t know what applications would be feasible,” says
Cerf.
Cerf and Kahn also couldn’t anticipate the
dangers the net would face in the 21st century. “The benefit of [the internet]
is that voices that might never have been heard are heard,” he tells me. “On
the other hand, we also hear from people whose messages we wish weren’t
delivered—terrorism, misinformation, deliberate deception, and then there is
malware and other kinds of things.”
Cerf, at 74, is still focused on the
future—including ensuring that future generations can read, watch, listen to,
and interact with digital creations of our time. Here are more of his thoughts,
edited for length and clarity, on a broad range of challenges facing the online
world.
PERSONAL RESPONSIBILITY ONLINE
Now we’re facing the question: What sort of
societal norms ought we invoke about behavior in the online world in order to
create a society that we would want to live in?
I’ve always interpreted the social contract
as the willingness of a citizen to curtail behavior in exchange for safety of
society and stability for society. And at this stage in the game, we have an
online society, which is still rather unmoved. And there are side effects that
are societally rather significant. The reported Russian intervention in the
election is just one example.
EDUCATING TECHNOLOGY USERS
My biggest concern is to equip the online
netizen with tools to protect himself or herself, to detect attempts to attack
or otherwise harm someone.
The term “digital literacy” is often referred
to as if you can use a spreadsheet or a text editor. But I think digital
literacy is closer to looking both ways before you cross the street. It’s a
warning to think about what you’re seeing, what you’re hearing, what you’re
doing, and thinking critically about what to accept and reject . . . Because in
the absence of this kind of critical thinking, it’s easy to see how the
phenomena that we’re just now labeling fake news, alternative facts [can come
about]. These [problems] are showing up, and they’re reinforced in social
media.
THE PROFUSION OF CONNECTED
GADGETS
What are the criteria that we should apply to
devices that are animated by software, and which we rely upon without
intervention? And this is the point where autonomous software becomes a concern,
because we turn over functionality to a piece of code. And dramatic examples of
that are self-driving cars . . . Basically you’re relying on software doing the
right things, and if it doesn’t do the right thing, you have very little to say
about it.
I feel like we’re moving into a kind of
fragile future right now that we should be much more thoughtful about
improving, that is to say making more robust.
THE RESPONSIBILITY OF
PROGRAMMERS
As software becomes more embedded in
everything that we do and rely on, the developers . . . will probably have to
accept more responsibility for the way in which they function . . . So I’m
expecting after various mishaps occur, some of which will lead to court cases,
there will probably be an evolving view of software responsibility.
What I’d like to do is get ahead of the
problem by getting programmers to feel a responsibility for making software
that is robust, that arguably . . . preserves privacy, that resists tampering,
protects against attack, and [provides] safety.
THE RELIABILITY OF THE INTERNET
Imagine a house that stops working when the
internet connection goes away. That’s not acceptable. So software has to work
autonomously, even when the internet is not there. And a lot of what falls into
the internet of things rubric is not designed with that in mind. There’s a
twisted assumption that the internet is [always going to be] there.
THE LIMITS OF ARTIFICIAL
INTELLIGENCE
Even playing a brilliant game of Go [a complex
Chinese board game that Google’s sister company DeepMind
recently mastered] is not the same as planning a vacation, driving
in a car across town, figuring out how to do your taxes. There’s just a huge
world out there that is not addressed by machine learning and artificial
intelligence.
When you experience the world, you create
models of it, and then reason about the models. So, an example: We’re sitting
here in this room with a glass of water on the shelf. Your model of this is
that it’s a container that contains liquid. So your use of this is not based on
a specific glass and specific water. But it is a container of liquid, and you
know that you can manipulate it.
Very few computer programs that we speak of
as artificially intelligent have the ability to abstract models from the real
world and reason about those models.
THE “DIGITAL DARK AGE”
Our media for writing has gotten less and
less [durable] as time has gone on . . . and you can certainly say that for
digital content. We miss one of two things: Either we don’t have any readers to
read the digital medium . . . or the bits that we preserve don’t mean anything
anymore because the software that knew what they meant doesn’t run.
My big worry is that this digital dark age
will be a consequence of losing the ability to correctly read or interpret
digital content . . . And so we have to take specific steps to preserve that
stuff . . . It’s not easy to get old software to run, because it may have been
designed for a particular hardware platform that doesn’t run or exist anymore.
THE PERISHABLE CLOUD
People are invested in the cloud, and part of
that is that you don’t have any choice. That’s just the way products and
services are offered to you. So I look at this, and I wonder if the cloud
implementers will last 100 years, or 500 years, or even 20 years. And if they
don’t, what’s the right thing to prepare for that?
At Google, we have a policy called the Data
Liberation Plan, which roughly speaking, says if you put it into the Google
system you should be able to get it back out again . . . But we still haven’t
found, once you get the data out, what do you to with it, where do you put it,
and how do you interpret it?
Despite everything he’s seen and created, Cerf
is still enchanted by technology. Though he doesn’t plug his employer very
much, Cerf admits being “in love with” Google Docs, which allows people around
the world to work on the same document simultaneously. “So, they replicate
copies, and they do it in real time,” he says. “I mean, how do they keep that
in sync?!” His sense of wonder is palpable–and striking, given that Docs, like
so many of the tools we use, build upon technologies that Cerf himself
envisioned long before they transformed the world.
BY SEAN CAPTAIN
https://www.fastcompany.com/40437402/the-internets-future-is-more-fragile-than-ever-says-one-of-its-inventors?utm_source=mailchimp&utm_medium=email&utm_campaign=fcdaily-top&position=8&partner=newsletter&campaign_date=07072017
No comments:
Post a Comment