To: NYC Council – Committee on Technology
From: Noel Hidalgo, Executive Director of BetaNYC
Re: Oversight hearing of Local Law 49 of 2018 (Open Algorithms Law) &
Int 1806-2019 (aka ADS transparency) & Int 1447-2019 (aka data inventory)
Thursday, 22 January 2020
“We […] want to ensure that New York City leads the way in ethical algorithmic government. We want transparency around data tools, algorithms, artificial intelligence, and tracking. We want New York City to be the thought leader in smart, ethical, algorithmic government.” – Noel Hidalgo, 4 Jan 2016
First, I want to say Happy New Year. We are glad to see Chairman Holden’s enthusiasm to use technology for good. Congratulations on your chairmanship, and we look forward to many hearings to come.
From BetaNYC’s point of view, these bills represent two of three battles for government transparency. Underpinning technology systems is data. Automated Decision Systems (ADS) is a function of software that affects us all. The third is actual software code and its design process.
For us to have government for the people, by the people, for the digital era, we must have transparent government software. To that end, we want more open source code within government.
Here are our thoughts on the two bills at hand.
Int 1806-2019 (aka ADS transparency)
We support the bill and suggest that the bill adopts the definition as specified by the AI Now report “Confronting Black Boxes – A Shadow Report of the New York City Automated Decision System Task Force.”
An “automated decision system” is any software, system, or process that aims to automate, aid, or replace human decision-making. Automated decision systems can include both tools that analyze datasets to generate scores, predictions, classifications, or some recommended action(s) that are used by agencies to make decisions that impact human welfare, and the set of processes involved in implementing those tools.
https://ainowinstitute.org/ads-shadowreport-2019.pdf
Int 1447-2019 (aka data inventory)
We support the bill but it needs significant modifications and conversations to ensure sustainability.
Reporting Deadline
The bill’s reporting date needs to better align with existing Mayor’s Office of Data Analytics (MODA) reporting. MODA is already incorporating portions of the old examination and verification (E&V) process into their annual open data report, which has historically come before this committee in the fall. Moving this bill’s reporting deadline brings about a natural alignment of existing practices. Also, we would like to note that this committee has NOT held its annual open data oversight hearing.
Scaling
The outcome of this bill should produce a sustainable, recursive process. To go deeper than the previous E&V process, there needs to be a learning process and alignment with existing open data reporting. Since we hope to be going deeper, we envision a process akin to the previous E&V — a process where a handful of agencies go through a data discovery process — then, learnings are quickly scaled to other agency audits. Since NYC Emergency Management already engages in a continuity of operations planning, agencies should be aware of their own technology systems and data sets that have to be maintained in two locations. You could even start with continuity of operations reports and publish a listing of those datasets and tools.
Our friends at GovEx Labs has a guide on how to marry data inventories with priorities and goals.
Accountability
If one of the goals is to get to a clear listing of data systems held at each agency, then we believe that MODA should be in charge of producing this report; however, Council Members need to hold agency leaders accountable. Over the past eight years, we have seen several data driven agencies resist posting data on the city’s open data portal. Then, in front of Council, see MODA interrogated for poor leadership at an Agency.
We want this bill to hold Agencies accountable.
As part of the last examination and verification report, the Department of Transportation (DOT) indicated it would post an additional 85 data sets on or before 31 December 2019 on the City’s open data portal. As of 21 January 2020, 45.88% (39) were posted or could be hiding under an existing dataset — over half (54% – 46 data sets) are not on the open data portal. I point to DOT as an example agency who has hosted several “open data sessions” with their CIO emphatically saying “if we have the data and if you want the data, we will get you the data.”
Additionally, this bill should state if the data is on the open data portal or a related dataset is on the City’s open data portal. If not on the open data portal, the report should state why it can not be posted to the open data portal.
Reviving the City’s Data Directory
Since the Charter revision of 1989, the Commission on Public Information and Communication (COPIC) has been in charge of producing an illustrious white whale. In April 1993, the first and only Data Directory was printed. Twenty seven years later, none of the open data / open government bills have ever gotten us to what was published in 1993.
Per the original data directory, the public was given a listing of:
- City databases which “contain information relating to the regulatory functions or statutory duties of an agency. Databases which are used for agency administrative support functions where not included (example accounting systems, personnel records, equipment inventory systems.”
- This includes: Agency contact information, Public Liaison contact information, Agency mission statement, Application Name, Year activated, Application description, Database contents
While this bill addresses data inventories, and not systems inventories, we believe that a comprehensive list of technology systems should be publicly available. If those systems were produced by the City, then their code should be available for public inspection.
If we say “in code we trust,” we must be able to see the code — whether it is law, software, or algorithms — we must have digital government transparency.
Future of Government Information and references to paper data…
The inventorying of paper data verges on the mission of NYC Records & Information Services. It is not clear how MODA is best suited to inventory data on paper. This bill has pointed out that freedom of information, automated decision making, and emergency management planning, all branch off from a simple and clear understanding of data and systems.
We look forward to these modifications and further discussion around digital transparency in New York City government.
“Our destiny is largely in our hands.” ― Frederick Douglass