How Policymakers Can Foster Algorithmic Accountability Joshua a - - PowerPoint PPT Presentation

THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

How Policymakers Can Foster Algorithmic Accountability

Joshua a New @josh sh_a _a_ne new Accountab abili lity y in the Algorithmic Economy May 22, 2018

THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

1. Algorithms pose new challenges 2. Existing proposals are flawed 3. Algorithmic accountability is the right approach 4. Implementing algorithmic accountability 5. Impact 6. Additional steps

OV OVERV RVIEW

THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

Complexity:

• Many ways bias can

influence an algorithm

• Difficult to interpret

Scalability:

• Risk amplifying flaws on

a large scale

AL ALGORITHMS POSE E NEW EW CHAL ALLENGES ES

Model of a neural network. Source: TeXample.net.

THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

• Hold algorithms to a standard

that does not exist for humans.

• Incentivize the use of less

effective AI.

• Assume the public and regulators

could interpret source code.

• Are useful in select contexts;

ineffective or harmful in most

• thers.

EX EXISTING PROPOSALS AL ALGORITHMS AR ARE E FLAWED WED

Angela Merkel discussing algorithmic

• transparency. Source: Medientage.

Mandates for algorithmic transparency and explainability:

THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

Master regulatory bodies:

• Ignores the need for

context-specific expertise.

• Assumes regulators cannot

develop the expertise to understand algorithms.

EX EXISTING PROPOSALS AL ALGORITHMS AR ARE E FLAWED WED

Elon Musk at the NGA 2017 Summer Meeting. Source: National Governors Association.

THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

Generalized regulatory proposals:

• Lack specifics about how to
• perationalize.
• Rely on platitudes that do not

translate to effective governance.

EX EXISTING PROPOSALS AL ALGORITHMS AR ARE E FLAWED WED

Theresa May at Davos 2018. Source: Number 10.

THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

Doing nothing:

• Market forces usually provide

adequate incentives:

• Bad decisions hurt a

company

• Consumer feedback and
• utrage
• Harms are minimal in many

cases

• Some use-cases are less subject

to these feedback mechanisms.

EX EXISTING PROPOSALS AL ALGORITHMS AR ARE E FLAWED WED

A ProPublica investigation revealing racial bias in COMPAS, a risk-assessment algorithm. Source: ProPublica.

THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

Algorithmic accountability is the principle that an algorithmic system should employ a variety of controls to ensure the “operator” (i.e., the party responsible for deploying the algorithm) can:

• Verify it acts in accordance with the
• perator’s intentions; and
• Identify and rectify harmful
• utcomes.

ALGORITHMI MIC ACCOU OUNTABILITY IS THE HE RI RIGHT HT APPR PPROACH

Pepper the robot. Source:Tokumeigakarinoaoshima.

THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

Verify it acts in accordance with the

• perator’s intentions:
• Transparency
• Explainability
• Confidence measures
• Procedural regularity

Identify and rectify harmful outcomes:

• Impact assessment
• Error analysis

DEFINING ALGORITHMIC C ACCOUNTABILITY

Datumbox Machine Learning Framework. Source: DatumBox.

THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

IMPLEME MENTING A ALGORI ORITHMI HMIC ACCOU OUNTABILITY

Was there unfair consumer injury? YES NO Did the operator have sufficient controls to verify its algorithm worked as intended? YES NO Did the operator identify and rectify harmful outcomes? YES NO Did the operator identify and rectify harmful outcomes? YES NO Low or no penalty Medium penalty High penalty No penalty

THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

IMPA PACT

• Operators have a clear understanding of

regulatory oversight and would proactively embrace algorithmic accountability.

• Market forces would encourage adherence to

algorithmic accountability.

THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

ADDI DDITIONAL S STE TEPS

• Adopt this as the U.S. approach

and advocate for its adoption abroad.

• Implement specific statutes for

algorithmic accountability for specific applications when appropriate.

• Increase regulators’ technical

expertise.

• Invest in new methods for

achieving algorithmic accountability.

Federal Trade Commissioner Joseph Simons. Source: Andrew Harrer/Bloomberg.

THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

CONCL NCLUSION

• Algorithms pose new

challenges, but existing proposals and the EU’s approach would not be effective and would harm innovation.

• Algorithmic accountability is

the right solution to the challenges posed by algorithmic decision- making.

European Commission. Source: Pixabay/Jai79.

THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

THA HANK YOU OU

• How Policymakers Can Foster Algorithmic Accountability:
• http://www.datainnovation.org/2018/05/how-policymakers-can-foster-

algorithmic-accountability/

• Email me: jnew@datainnovation.org
• @josh_a_new