Comprehensive software reviews to make better IT decisions
AI Registers: Finally, a Tool to Increase Transparency in AI/ML
Transparency, explainability, and trust are pressing issues in artificial intelligence (AI) and machine leaning (ML) today. Nobody wants to find themselves at the receiving end of a black box system that makes consequential decisions about (for example) jobs, healthcare, and citizenship, especially if those decisions are unfair, biased, or just not in one’s favor. And most organizations agree that consumer trust and confidence in AI being used ethically and transparently are key to unlocking its true potential.
And while there are literally hundreds of documents describing and prescribing AI principles, frameworks, and other good guidelines, there haven’t been any practical tools that could help with implementing transparency. Until now.
On September 28, 2020, the cities of Helsinki and Amsterdam jointly announced the launch of their public AI registers. The two registers were developed in collaboration with Saidot.ai, an innovative Finnish company which specializes in bringing transparency to consumer services and which, to my knowledge, is the only vendor in this space.
The idea for the company grew from the personal frustration of Saidot’s founder and CEO Meeri Haataja, who was “seeing how important transparency of AI is for the future of each one of us, and not being able to find too many meaningful ways to act on it.”
What Is an AI Register?
The City of Helsinki defines an AI register as “a window into the artificial intelligence systems used by the City”. The City of Amsterdam calls theirs an “algorithm register” and describes it as “an overview of the artificial intelligence systems and algorithms used by the City of Amsterdam.” Screenshots of both of these applications are found below.
Both registers describe:
- Where and how their respective cities are using AI/ML.
- How these applications impact citizens and their everyday lives.
- How they were built.
- Which data and algorithms they use.
- What kinds of decisions and assumptions were made in the design, build, and deployment of the apps.
- Which ethical principles were employed to mitigate biases and other risks.
The registers also include contact information for further inquiries or to provide feedback, as seen in this screenshot from the City of Amsterdam:
The language used in these AI registers is clear and easy to understand, aimed at average citizens without any technical background in AI/ML. (Some technical information is provided as well; while it is limited, one could probably obtain more via the contacts mentioned above.)
So far only a small number of AI applications have been documented in the city registers – five by the City of Helsinki and four by the City of Amsterdam – but more are in the works.
Benefits of AI Registers for Governments and Citizens
By disclosing information about their apps, governments improve transparency and accountability around their use of AI. This helps to increase citizen trust, and trust is the foundation of any government. “The government can only function when there is trust between the government and people,” says this whitepaper from Saidot.ai. Or, as the City of Helsinki’s Chief Digital Officer Mikko Rusama says, “Without trust, there is no use for AI.”
For citizens, AI registers help to increase their AI/ML literacy. They promote citizens’ awareness of and participation in public debate – one of the cornerstones of democratic governments. They build trust by increasing transparency in how, where, and why the government uses their data, taxpayer money, and new technologies such AI/ML.
Ultimately, what we all want is responsible AI benefits all of us, respects diversity and human rights, encourages human flourishing, sustains human agency, and promotes societal, economic, and environmental sustainability.
Benefits of AI Registers for AI Teams
For AI teams, AI registers offer a mechanism to “standardize transparency across your AI portfolio” – a “searchable and archivable way to document the decisions and assumptions that were made in the process of developing, implementing, managing, and ultimately dismantling an algorithm,” according to this whitepaper from Saidot.ai. Saidot’s software comes with“pre-defined modular metadata models which are adaptable to organization and sector-specific requirements, while securing interoperability between different versions.” Its offering consists of a backend platform, the register itself, and an interface for publishing the registers (which the company hosts).
Source: Screenshot from What forms of mandatory reporting can help achieve public-sector algorithmic accountability? A virtual event from the Ada Lovelace Institute featuring speakers Soizic Penicaud, Meeri Haataja, Natalia Domagala, and Matthias Spielkamp. [CC-BY 4.0]
In addition to the cities of Amsterdam and Helsinki, the company lists Finnair, the City of Espoo, as well as Finland’s Ministry of Finance, Social Insurance Institute, and Ministry of Justice among its customers and partners.
Saidot continues to actively develop its platform. New features and directions it’s currently considering include:
- Integrating AI registers into their clients’ AI development processes, which would help expand it beyond communications and into solution governance, solidifying accountability.
- Using AI registers to capture citizen feedback early in the design and ideation process, to confirm desirability and enable co-design and co-creation.
- Facilitating information sharing between government organizations to help with, for example, vendor selection. (The registers disclose whether external vendors were engaged and who they are.)
Beyond Government: Why You Should Adopt AI Registers for Your AI/ML and Data Science Apps
Trust is not just a fundamental prerequisite for a government to exist. Trust is equally fundamental in business. The entire reason your organization exists is because of your customers, consumers, employees, and partners. Whether you are just getting your feet wet with AI/ML or are running your whole business on a foundation of AI/ML, follow Amsterdam’s and Helsinki’s examples. Document, make public, and/or share with employees and partners what your organization is using AI/ML and data science for and how. We’ll all benefit from this disclosure. And the AI register software from Saidot has just made implementing transparency much, much easier.
Want to Know More?
To learn more about AI registers and Saidot’s platform:
- Read this short whitepaper Public AI Registers: Realizing AI transparency and civic participation in government use of AI.
- Check out the City of Amsterdam’s Algorithm Register and the City of Helsinki’s AI Register.
- Contact Saidot to join their beta program.
- If your organization is part of the public sector, you can get free access to Saidot’s backend system for one year. (The offer is valid until December 31, 2020 as per Info-Tech's conversation with Saidot's CEO.)
Recently I attended the inaugural Emotion AI conference, organized by Seth Grimes, a leading analyst and business consultant in the areas of natural language processing, text analytics, sentiment analysis, and their business applications. So, what is emotion AI, why is it relevant, and what do you need to know about it?
SortSpoke’s novel approach to machine learning answers a longstanding problem in financial services – how to efficiently extract critical data from inbound, unstructured documents at 100% data quality.
Amazon is offering its cashierless store technology to other retailers. The technology known as “Just Walk Out” eliminates checkout lines, offering an “effortless” shopping experience and shifting store associates to “more valuable activities”.
As the COVID-19 pandemic is shutting down whole countries, a few of you may be wondering whether AI can help create a vaccine for the virus responsible. After all, AI is magic, right?
Alphabet is facing backlash from its shareholders over its approach to digital privacy, reports the Financial Times. And not for the first time. This time, however, things will need to change.
The EU plans to invest €6 billion to build a single European data space, reports EURACTIV. The envisioned space will house personal, business, and “high-quality industrial data” and create the infrastructure for data sharing and use across businesses and nations.
“Facebook quietly acquired another UK AI startup and almost no one noticed,” reported TechCrunch on February 10. We looked into why.
In a landmark ruling, a Dutch court has ordered an immediate halt to the government’s use of an automated system for detection of welfare fraud.
Databricks, a data processing and analytics platform with a strong focus on AI and ML, has partnered with Immuta to deliver automated end-to-end data governance for AI, data science, and ML projects.