AI Ethics #32: Vulnerability in the Gig Economy in Africa, Let's Talk Privacy, and bridging the gap between AI and the public
AI Ethics #32: Vulnerability in the Gig Economy in Africa, Let's Talk Privacy, and bridging the gap between AI and the public
Are you aware of the best practices for creating and using keywords to document indigenous artifacts?
Welcome to another edition of our weekly newsletter that will help you navigate the fast-changing world of AI Ethics! Every week we summarize the best of AI Ethics research and reporting, along with some commentary. More about us at montrealethics.ai/about.
Photo by Kai Pilger on Unsplash
This week’s overview:
Research summaries:
Between a Rock and a Hard Place: Freedom, Flexibility, Precarity and Vulnerability in the Gig Economy in Africa
Project Let’s Talk Privacy
Op-ed:
Bridging the Gap Between AI and the Public (TEDxYouth@GandyStreet)
Article summaries:
Best Practices for Indigenous Keywording for Stock Images (Shutterstock Blog)
Standing up for developers: youtube-dl is back (Github Blog)
How China surveils the world (MIT Technology Review)
Health Care AI Systems Are Biased (Scientific American)
Australia’s Artificial Intelligence (AI) future: A call to Action (Lexology)
Software is trying to change your habits. Make sure you agree with it. (Zapier Blog)
But first, our call-to-action of the week: The State of AI Ethics (panel discussion), hosted by us!
Following up on our last report, we decided to invite some of the contributors back for a virtual panel discussion. There will be time for audience Q&A!
Sign up here.
Topic: 2020 in review, from an AI Ethics perspective
Speakers: Rumman Chowdhury (Accenture), Danit Gal (United Nations), Katya Klinova (Partnership on AI), Amba Kak (NYU’s AI Now Institute), Abhishek Gupta (Montreal AI Ethics Institute). Moderated by Victoria Heath (Montreal AI Ethics Institute).
Date: Wednesday, December 2nd from 12:30 PM EST – 2:00 PM EST
Free tickets via Eventbrite: here!
Research summaries:
Between a Rock and a Hard Place: Freedom, Flexibility, Precarity and Vulnerability in the Gig Economy in Africa by Mohammad Amir Anwar, Mark Graham
This paper lays out the on-the-ground realities of freelancing digital workers in Africa and points to how the growth of the digital gig economy is reinforcing the asymmetrical global division of labour.
To delve deeper, read our full summary here.
Project Let’s Talk Privacy by Anna Chung, Dennis Jen, Jasmine McNealy, Pardis Emami Naeni, Stephanie Nguyen
This project provides some empirical evidence from a design thinking perspective in terms of how privacy policies and bills associated with the passing of different privacy legislations ought to be designed so that they best communicate their message to the intended audiences.
To delve deeper, read our full summary here.
What we’re thinking:
Bridging the Gap Between AI and the Public (TEDxYouth@GandyStreet) by Connor Wright
With the theme of “bridging the gap”, I decided to base my TEDx Youth talk on bridging the gap between the public and the AI debate. Given that AI is often thought of as something reserved for killer robots, I wanted to show how AI in its current format could potentially achieve far worse than a killer robot ever could. To do this, I presented AI in the form of algorithms being applied to different aspects of human life, before launching into my argument. Here, after I compared the current AI situation to a novel in progress, I mentioned two negative consequences of the public not getting involved: a lack of diversity in the AI building process and a lack of pushback. I’ll now walk you through how this took shape.
To delve deeper, read the full article here.
Article summaries:
Best Practices for Indigenous Keywording for Stock Images (Shutterstock Blog)
One of the best articles to have come to our attention in recent times because it provides clear guidance on how language creates power dynamics and what we can do to avoid intentionally and unintentionally weaponizing language. The article details the Shutterstock team’s efforts in including more inclusive and respectful language for the images associated with Indigenous Peoples. The article provides great detail in terms of the phrasing to use, the capitalization, punctuation, and other language specifics so that we respect the wishes of the people we are trying to represent.
One thing that we really liked was how it paid heed to geographical differences and provided clear guidelines in terms of phrases that are appropriate to use and those that are not. In cases where there might be ambiguity, the golden rule of asking those you are trying to represent about the phrasing that they would like to see used about images or any other assets is the best way to go.
While not reproducing the list here (we strongly encourage you to read the article in case you are interested), what we found relevant for our discussions is the impact that such inclusive labeling will have downstream on systems that scrape metadata to train machine learning systems. Also, this has impacts from an archival perspective in terms of what future generations will have as artifacts when they seek to understand the state of a community in a bygone era. Our team presented some work earlier this year with the title “Comprehensiveness of Archives: A Modern AI-enabled Approach to Build Comprehensive Shared Cultural Heritage” that touched on these ideas in terms of representation in archives.
Standing up for developers: youtube-dl is back (Github Blog)
(Full disclosure: Our founder Abhishek Gupta works at Microsoft, which owns GitHub. However, the inclusion of this article in the newsletter is unrelated to his employment and not paid for or endorsed by Microsoft)
If you’re not familiar with the youtube-dl tool, it is a nifty utility that helps you download videos from YouTube. The immediate concern that comes to mind is that it might be misused to gather copyrighted content right? Not entirely. There are many legitimate uses of such tools, say for fair use, demonstrating harm, gathering evidence, documentation on the part of journalists, downloading public-domain videos, changing playback speeds for accessibility, amongst other uses. So, the internet was in an uproar when GitHub decided to take down the popular repository in response to some DMCA notices (a notice from copyright owners when they believe there has been an infringement). This is not atypical for any website that hosts content generated by users.
So, this article provided much-needed clarity on how this particular request was handled, why it was different from others that they have received, and the new, more robust practices that are being instituted by GitHub to prevent unnecessary takedowns. The repository was finally restored after GitHub deemed that there weren’t adequate grounds for the removal of the repository.
The crux of the argument from the people that requested the takedown of the repository was that the tools allowed the circumvention of technical protection measures (TPM) that are supposed to protect the content and rights holders as enshrined in Section 1201 of the DMCA. The original formulation of the DMCA from the 90s doesn’t account for cases where such circumvention might not lead to copyright infringement and hence limits all such software.
In a nutshell, after the maintainer of the repository made the requisite changes around how some of the unit tests in the repository were utilizing copyrighted videos, and keeping in line with the ethos of supporting developers, GitHub reinstated the repository. Going forward, they have also set up more robust guidelines so that such incidents are minimized, and they will lean towards erring on the side of developers.
How China surveils the world (MIT Technology Review)
A little bit akin to gaining an inside hold on the infrastructure being built in countries through massive aid provided to emerging economies and the rapid expansion of the Belt and Road project from China, the use of partnerships with universities, access to data from social media platforms, and apps that store data in the PRC provide CCP with a convenient mechanism to gather data on the entire world.
As pointed out in the article, the collection of data sometimes doesn’t carry any immediate purpose but is being done proactively so that it might be used in the future, perhaps gleaning insights that aren’t possible with technical solutions today.
The storage location of the data is also a hot-button issue, because of the ability of the CCP to requisition data as it sees fit, something that was a sticking point in the TikTok bans in several countries and the fiasco with trying to find US buyers to avoid this very concern amongst others. The article also mentions a company GTCOM, similar to Palantir, that ingests a lot of data and provides intelligence from that to various actors. The kind of data that is utilized runs the gamut: voice, text, images, video, metadata, and everything else. Given that the agenda for this data gathering is ushered by the CCP, this is an area of severe concern as it wrests a great deal of power in the hands of an authoritarian government concerning people from around the world.
Health Care AI Systems Are Biased (Scientific American)
Bias in AI is a topic that we have covered numerous times in this newsletter. This article elucidates the problem by talking about the highly fragmented nature of the data ecosystem in the healthcare system. Rightly so there are concerns about privacy given the sensitive nature of the information that is stored in these databases. But, a side effect of this is that each institution trying to use techniques that require large amounts of data struggle because they are unable to compile together enough data that can form representative samples and hence open up the possibility of bias.
The use of techniques like differential privacy can potentially help to mitigate some of these concerns, but there are vested interests by some healthcare providers who want to keep the data sequestered in their silos for the fear of losing business. A study mentioned in the article talks about how healthcare institutions that are more open with their data see higher rates of patient attrition compared to those that don’t.
Perhaps, there is also risk mitigation and aversion that comes into play whereby administrators don’t want to take on potential problems where data sharing can lead to adverse outcomes from a legal and compliance perspective. National repositories and data commons structured through trusts can be a way around some of these problems, but they require concerted efforts on the part of multiple parties before they become a reality. Most importantly, we believe that awareness on the part of patients so that they don’t just accept the status quo but see data as an extension of their self can also help to push this agenda and ensure that there is more openness in sharing data and hence better chances at mitigating bias in AI systems used in healthcare.
Australia’s Artificial Intelligence (AI) future: A call to Action (Lexology)
Good to see how Australia has made access to high-quality data that is desensitized and ethically and legally sourced as a core tenet within their Action Plan for the adoption and use of AI. Especially encouraging is the fact that practitioners and researchers are going to be working together on addressing these issues. Without such collaboration between the theoretical and practical, a lot of the initiatives fail to meet their targets.
One of the shortcomings though identified in the article is the inadequacy of the support from a financial perspective which is also a key component to the success of any such work. Numerous self-funded initiatives peter out over time because of funds drying up and volunteers being forced to make a choice between sharing their expertise without compensation and being fiscally viable at a personal level.
Something that caught the eye in this article was the emphasis on having “AI translators”, people who are able to navigate different domains so that the capabilities and limitations of these systems can be adequately communicated. Often, we are overzealous in our expectations with what AI-enabled systems can do which leads to unfortunate outcomes from an ethical standpoint. Coupling the upskilling of existing staff to take advantage of AI advances and providing clear guidance for businesses on the ethical considerations will be critical for the success of a responsible AI ecosystem.
Software is trying to change your habits. Make sure you agree with it. (Zapier Blog)
While ultimately the article does talk about how Zapier might be used in a way that helps us break out of some habits, the thrust of the article is quite relevant to discussions on the societal impacts of technology and the role that design plays in that. Drawing on The Gruen Effect and talking about how malls are intentionally designed to be all-encompassing and labyrinth-like making it harder for us to stick to our goals, the article makes a strong point on the same thing that happens with the apps that we use.
When notifications on Facebook initially started off as a way to see if anyone had tagged you in something, commented on your stuff, or liked something that you had shared, we now get notifications for unrelated activity in the hopes that when we open the website/app, we find something else that catches our attention and we end up spending our precious time. While these notifications are tunable to prevent this kind of behaviour, rarely do we spend time doing that. While we might not consciously spend too much time thinking about these design decisions, the companies spend an inordinate amount of time doing so to harvest our attention.
Some habits are good and we might want to have the apps guide us in sticking to our goals of eating healthy and exercising more, but there are many more negative habits that the apps help to enforce and knowing the agenda of the apps and what tactics might be used to hold your attention in unwanted ways can help us make better decisions.
From elsewhere on the web:
Overview of The State of AI Ethics (October 2020) (INDIAai)
INDIAai summarizes our report, explains its relevance within a broader context, and provides some quick takeaways.
Annual Digital Open Government Forum: Harnessing the Power of Artificial Intelligence (Government of Canada)
Our researcher Erick Galinkin will be doing an ask-me-anything on the topic of building transparency, accountability and integrity in AI.
Guest post:
If you’ve got an informed opinion on the impact of AI on society, consider writing a guest post for our community — just send your pitch to support@montrealethics.ai. You can pitch us an idea before you write, or a completed draft.
In case you missed it:
The Short Anthropological Guide to the Study of Ethical AI
This short guide serves as both an introduction to AI ethics and social science and anthropological perspectives on the development of AI. It intends to provide those unfamiliar with the field with an insight into the societal impact of AI systems and how, in turn, these systems can lead us to rethink how our world operates.
To delve deeper, read the full piece here.
Take Action:
Community Nomination Form - State of AI Ethics January 2021 Report
We’re inviting the AI ethics community to nominate researchers, practitioners, advocates, and community members in the domain of AI ethics to be featured in our upcoming State of AI Ethics report.
MAIEI Learning Community
Interested in discussing the ethical challenges of AI in addressing some of the biggest ethical challenges of AI to develop interdisciplinary solutions with thinkers from across the world?
Our AI Ethics consulting services
In today’s market, the make-or-break feature for organizations using AI is whether they embody the principles of morality and ethics.
We want to help you analyze your organization and/or product offerings for moral and ethical weaknesses and make sure your business model is airtight. By undergoing a thorough, multidisciplinary review of your AI tool, we will provide ethical, social, and technical feedback on your work and research, which will allow you to proactively address your own blindspots and maximize your potential before ever undergoing a third-party ethics review.
Events:
As a part of our public competence building efforts, we host events frequently spanning different subjects as it relates to building responsible AI systems. We also share events from the broader AI ethics ecosystem.
The State of AI Ethics (Panel), hosted by us!
Topic: 2020 in review, from an AI Ethics perspective
Speakers: Rumman Chowdhury (Accenture), Danit Gal (United Nations), Katya Klinova (Partnership on AI), Amba Kak (NYU’s AI Now Institute), Abhishek Gupta (Montreal AI Ethics Institute). Moderated by Victoria Heath (Montreal AI Ethics Institute).
Date: Wednesday, December 2nd from 12:30 PM EST – 2:00 PM EST
Free tickets via Eventbrite: here!
Signing off for this week, we look forward to it again in a week! If you enjoyed this and know someone else that can benefit from this newsletter, please share it with them!
If someone has forwarded this to you and you want to get one delivered to you every week, you can subscribe to receive this newsletter by clicking below:
If you have feedback for this newsletter or think there’s something interesting we missed, email us at support@montrealethics.ai