Privacy versus government surveillance – where
MS
Published · 23 slides · 0 views
1 / 1
Description
Privacy versus government surveillance where network effects meet public choice Ross Anderson Cambridge Berkeley Law School Oct 6 2014 Two views of money and power The Bay Area view: money and power are all about network effects, which
Related Topics
Share
Embed code
Download this presentation From Below
"Privacy versus government surveillance – where" is the property of its rightful owner. Permission is granted to download and print the materials on this website for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Presentation Transcript
01
Privacy versus government surveillance – where network effects meet public choice Ross Anderson
Cambridge Berkeley Law School Oct 6 2014<br>
Cambridge Berkeley Law School Oct 6 2014<br>
02
Two views of money and power The Bay Area view: money and power are all about network effects, which help you create a platform to which everyone else then adds value
The Washington DC view: power is about having more tanks and aircraft carriers, which is founded on taxation capacity
Almost no-one talks of network effects there, or among scholars of government! Berkeley Law School Oct 6 2014<br>
The Washington DC view: power is about having more tanks and aircraft carriers, which is founded on taxation capacity
Almost no-one talks of network effects there, or among scholars of government! Berkeley Law School Oct 6 2014<br>
03
Is this changing? 1980s: a non-aligned country like India is a democracy, but buys its jet fighters from Russia because they’re cheaper
2000s: Snowden tells us that India shares intelligence with the NSA rather than the FSB, as the NSA’s network is bigger
The “five eyes” is maybe 15 eyes, or 35 eyes, or 65 eyes … Berkeley Law School Oct 6 2014<br>
2000s: Snowden tells us that India shares intelligence with the NSA rather than the FSB, as the NSA’s network is bigger
The “five eyes” is maybe 15 eyes, or 35 eyes, or 65 eyes … Berkeley Law School Oct 6 2014<br>
04
Information economics (1) The first characteristic of many IT product and service markets is network effects
Metcalfe’s law – the value of a network is the square of the number of users
Real networks – phones, fax, email, Facebook, Visa/Mastercard
Virtual networks – PC versus Mac, Visa/MC
Network effects tend to lead to dominant-firm markets where the winner takes all Berkeley Law School Oct 6 2014<br>
Metcalfe’s law – the value of a network is the square of the number of users
Real networks – phones, fax, email, Facebook, Visa/Mastercard
Virtual networks – PC versus Mac, Visa/MC
Network effects tend to lead to dominant-firm markets where the winner takes all Berkeley Law School Oct 6 2014<br>
05
Information economics (2) Second common feature of IT product and service markets is high fixed costs and low marginal costs
Competition can drive down prices to marginal cost of production
This can make it hard to recover capital investment, unless stopped by patent, brand, network effects …
These effects can also lead to market power Berkeley Law School Oct 6 2014<br>
Competition can drive down prices to marginal cost of production
This can make it hard to recover capital investment, unless stopped by patent, brand, network effects …
These effects can also lead to market power Berkeley Law School Oct 6 2014<br>
06
Information economics (3) Third common feature of IT markets is that switching is expensive
E.g. switching from Windows to Linux means retraining staff, rewriting apps
And once you have $3000 worth of songs on a $300 iPod, you’re locked into iPods etc
Shapiro-Varian observation: the net present value of a software company is the total switching costs Berkeley Law School Oct 6 2014<br>
E.g. switching from Windows to Linux means retraining staff, rewriting apps
And once you have $3000 worth of songs on a $300 iPod, you’re locked into iPods etc
Shapiro-Varian observation: the net present value of a software company is the total switching costs Berkeley Law School Oct 6 2014<br>
07
Economics of (information) security Each of network effects, low marginal costs and technical lock-in makes dominant-firm market structures more likely
Together, they make them much more likely
They also explain many security and privacy failures
First, market races lead to “Ship it Tuesday and get it right by version 3” Berkeley Law School Oct 6 2014<br>
Together, they make them much more likely
They also explain many security and privacy failures
First, market races lead to “Ship it Tuesday and get it right by version 3” Berkeley Law School Oct 6 2014<br>
08
Security economics (2) In a market race, you open your system to appeal to complementers such as app writers
Once you’ve won, you lock it down to extract rents
In one market after another – mainframes, PCs, routers, phones, social network systems – security is added later
Its design ends up aligned with the platform’s interests at least as much as the users’ Berkeley Law School Oct 6 2014<br>
Once you’ve won, you lock it down to extract rents
In one market after another – mainframes, PCs, routers, phones, social network systems – security is added later
Its design ends up aligned with the platform’s interests at least as much as the users’ Berkeley Law School Oct 6 2014<br>
09
Economics of privacy Also a Berkeley thing (Alessandro Acquisti, Jens Grossklags, Hal Varian …)
Privacy suffers from the same problems as security, and more
Asymmetric information: users don’t know much about what gets done with their data
Hyperbolic discounting: many users don’t care about long-term effects of disclosure
People say they want privacy but act otherwise (privacy startups tend to go bust) Berkeley Law School Oct 6 2014<br>
Privacy suffers from the same problems as security, and more
Asymmetric information: users don’t know much about what gets done with their data
Hyperbolic discounting: many users don’t care about long-term effects of disclosure
People say they want privacy but act otherwise (privacy startups tend to go bust) Berkeley Law School Oct 6 2014<br>
10
Information security economics Since the first workshop on the economics of information security, here at Berkeley in 2002, it’s grown to over 100 active researchers
Models of what’s likely to go wrong – perverse incentives, asymmetric information
Measurements of what is going wrong – patching cycle, malware, fraud
Recommendations – what actors can likely do what
Policy recommendations now being adopted in both the USA and Europe (but often twisted by lobbyists)
Extending into behavioral economics, psychology Berkeley Law School Oct 6 2014<br>
Models of what’s likely to go wrong – perverse incentives, asymmetric information
Measurements of what is going wrong – patching cycle, malware, fraud
Recommendations – what actors can likely do what
Policy recommendations now being adopted in both the USA and Europe (but often twisted by lobbyists)
Extending into behavioral economics, psychology Berkeley Law School Oct 6 2014<br>
11
Next – economics of surveillance? Network effects brought us mass surveiilance!
The concentration of the industry into a few large service firms (MS, G, Y, FB …) made the PRISM program foreseeable (except in its details)
The concentration of the telecomms industry into a handful of large operators similarly made TEMPORA foreseeable (it was described by several journalists in its earlier form of ‘Echelon’)
But that’s not all! Berkeley Law School Oct 6 2014<br>
The concentration of the industry into a few large service firms (MS, G, Y, FB …) made the PRISM program foreseeable (except in its details)
The concentration of the telecomms industry into a handful of large operators similarly made TEMPORA foreseeable (it was described by several journalists in its earlier form of ‘Echelon’)
But that’s not all! Berkeley Law School Oct 6 2014<br>
12
Information economics and defence (1) Network effects do matter in the defence / intelligence nexus!
Neutrals like India prefer to join the biggest network
Subnetworks feed things in too
OTOH, network effects entangle us with bad states which use the same surveillance platforms (see rows over exports to Syria) Berkeley Law School Oct 6 2014<br>
Neutrals like India prefer to join the biggest network
Subnetworks feed things in too
OTOH, network effects entangle us with bad states which use the same surveillance platforms (see rows over exports to Syria) Berkeley Law School Oct 6 2014<br>
13
Information economics and defence (2) Medieval warfare was all run on marginal costs (40-60 days service for every peasant)
WW1: sent millions of men to Germany
WW2: fewer front-line troops, but lots of planes, tanks and other capital equipment
Now: to kill a foreign dictator you can use a single missile shot from a drone – because it’s backed by trillions of capital investment Berkeley Law School Oct 6 2014<br>
WW1: sent millions of men to Germany
WW2: fewer front-line troops, but lots of planes, tanks and other capital equipment
Now: to kill a foreign dictator you can use a single missile shot from a drone – because it’s backed by trillions of capital investment Berkeley Law School Oct 6 2014<br>
14
Information economics and defence (3) Complex technical lock-in games
1980s: it was basically about ammunition and spares
Now: are you using Cisco or Huawei?
Very expensive try to build independent infrastructure for government networks
Even so, shared code can lead to shared attacks Berkeley Law School Oct 6 2014<br>
1980s: it was basically about ammunition and spares
Now: are you using Cisco or Huawei?
Very expensive try to build independent infrastructure for government networks
Even so, shared code can lead to shared attacks Berkeley Law School Oct 6 2014<br>
15
Intelligence network governance Core is 5 eyes; expanding circles of others
Governance: each agency could decide whether to minimise its citizens’ personal data
Only Canada did so!
So GCHQ is happy for the NSA to read my medical records, and the NSA is happy for GCHQ to read yours!
The NSA was also quite happy for Aussies to read US attorney-client communications Berkeley Law School Oct 6 2014<br>
Governance: each agency could decide whether to minimise its citizens’ personal data
Only Canada did so!
So GCHQ is happy for the NSA to read my medical records, and the NSA is happy for GCHQ to read yours!
The NSA was also quite happy for Aussies to read US attorney-client communications Berkeley Law School Oct 6 2014<br>
16
Law enforcement network governance Various models from Interpol through mutual legal assistance treaties
Very slow and cautious: requests vetted by both governments, often several agencies
Much effort on accelerating the process, e.g. via personal links created from NCFTA training and exchange programs
In theory, everything is filtered by lawyers Berkeley Law School Oct 6 2014<br>
Very slow and cautious: requests vetted by both governments, often several agencies
Much effort on accelerating the process, e.g. via personal links created from NCFTA training and exchange programs
In theory, everything is filtered by lawyers Berkeley Law School Oct 6 2014<br>
17
One network or many? Networks tend to merge: the Internet absorbs everything else
Will the intelligence network and the law-enforcement network become one?
Already intelligence resources are used for rapid solution of exceptional crimes
UK: NTAC and the Communications Data Bill
USA: PRISM Berkeley Law School Oct 6 2014<br>
Will the intelligence network and the law-enforcement network become one?
Already intelligence resources are used for rapid solution of exceptional crimes
UK: NTAC and the Communications Data Bill
USA: PRISM Berkeley Law School Oct 6 2014<br>
18
Governance and networks PCAST report conceptualised ‘Big data’ as collectors v aggregators v users
It argued that only the users – the firms who buy ads – should be regulated
Within a week, the European Court of Justice disagreed in the Gonzales case
The aggregator can republish obscure facts, and also combine non-sensitive facts to make a sensitive one Berkeley Law School Oct 6 2014<br>
It argued that only the users – the firms who buy ads – should be regulated
Within a week, the European Court of Justice disagreed in the Gonzales case
The aggregator can republish obscure facts, and also combine non-sensitive facts to make a sensitive one Berkeley Law School Oct 6 2014<br>
19
Governance and networks (2) With a converged intelligence network, user controls might mean
NSA can use it all
FBI can use X
State police can use Z …
What price separation of powers?
What controls – legal or technical – are feasible for collection, and for aggregation?
If code is law, then architecture is policy… Berkeley Law School Oct 6 2014<br>
NSA can use it all
FBI can use X
State police can use Z …
What price separation of powers?
What controls – legal or technical – are feasible for collection, and for aggregation?
If code is law, then architecture is policy… Berkeley Law School Oct 6 2014<br>
20
The IR Community Realists (Thucydides, Machiavelli, Hobbes, Kissinger …) vs idealists / liberals (Kant, Wilson, Keohane, Clinton …)
Not even the latter seem to have considered network effects (rare passing references only)
Yet network effects surely add weight to the liberal side of the argument
A realist interpretation of the surveillance world would have depressing implications … Berkeley Law School Oct 6 2014<br>
Not even the latter seem to have considered network effects (rare passing references only)
Yet network effects surely add weight to the liberal side of the argument
A realist interpretation of the surveillance world would have depressing implications … Berkeley Law School Oct 6 2014<br>
21
The long term Britain’s value to the USA: we provide access to 30% of the Internet
Thank you, Queen Victoria!
If code is law, architecture is policy
What are we embedding in the infrastructure, and how will it affect our descendants?
Another reason to favour the liberal view of IR
We need to solve the governance problem Berkeley Law School Oct 6 2014<br>
Thank you, Queen Victoria!
If code is law, architecture is policy
What are we embedding in the infrastructure, and how will it affect our descendants?
Another reason to favour the liberal view of IR
We need to solve the governance problem Berkeley Law School Oct 6 2014<br>
22
More network effects in government Example 1: the EU itself as a customs union, which imposes its laws de facto on neighbouring states (Norway, Iceland, Switzerland …)
Example 2: financial regulation – disproportionate leverage of US, UK
Example 3: the EU smart meter programme, which aimed at energy efficiency, but was fragmented by national energy markets
… Berkeley Law School Oct 6 2014<br>
Example 2: financial regulation – disproportionate leverage of US, UK
Example 3: the EU smart meter programme, which aimed at energy efficiency, but was fragmented by national energy markets
… Berkeley Law School Oct 6 2014<br>
23
Conclusions The SF-DC gap is not just whether Snowden’s a whistleblower or a traitor! The economic models are almost totally different
Yet models of the economics of security and privacy that were pioneered at Berkeley a dozen years ago can apply here too
The implications range from international relations to the separation of powers and the rule of law Berkeley Law School Oct 6 2014<br>
Yet models of the economics of security and privacy that were pioneered at Berkeley a dozen years ago can apply here too
The implications range from international relations to the separation of powers and the rule of law Berkeley Law School Oct 6 2014<br>