AI for Everyone

A new book on the ethical issues of Artificial Intelligence has been published by the University of Westminster.  I am thrilled it has been published under Open Access, so is free to any who want it.  This is how academic research should be done!

You can download a copy of this excellent book here, or from the University of Westminster website.  Please share it around!

AI for Everyone? Critical Perspectives

Edited by Pieter Verdegem

This book is distributed under the terms of the Creative Commons Attribution + Noncommercial + NoDerivatives 4.0 license. Copyright is retained by the author(s)

This book has been peer reviewed.

ABSTRACT

We are entering a new era of technological determinism and solutionism in which governments and business actors are seeking data-driven change, assuming that Artificial Intelligence is now inevitable and ubiquitous. But we have not even started asking the right questions, let alone developed an understanding of the consequences. Urgently needed is debate that asks and answers fundamental questions about power. This book brings together critical interrogations of what constitutes AI, its impact and its inequalities in order to offer an analysis of what it means for AI to deliver benefits for everyone. The book is structured in three parts: Part 1, AI: Humans vs. Machines, presents critical perspectives on human-machine dualism. Part 2, Discourses and Myths About AI, excavates metaphors and policies to ask normative questions about what is ‘desirable’ AI and what conditions make this possible. Part 3, AI Power and Inequalities, discusses how the implementation of AI creates important challenges that urgently need to be addressed. Bringing together scholars from diverse disciplinary backgrounds and regional contexts, this book offers a vital intervention on one of the most hyped concepts of our times.

DOWNLOAD AI FOR EVERYONE

ORIGINAL https://www.uwestminsterpress.co.uk/site/books/e/10.16997/book55/

Conference – Ethics, Human Rights & Emerging Technologies 10-12 March

Ethics, Human Rights & Emerging Technologies 10-12 March

Save the date for our final (online) event on Ethics, Human Rights & Emerging Technologies on March 10, 11 and 12! In the conferencewe will discuss ethical and human rights issues raised by emerging technologies, and will assess the new methods and instruments we need for ethical guidance and governance of emerging technologies.

Over three days, we will address instruments like regulation, innovation policies, research ethics frameworks, Ethics by Design methodologies, education and training progammes, standards, and certification. Here, we will present and discuss results of the 3,5 years of work that the SIENNA project has conducted. The conference consists of four parts, which can be attended separately.

Watch this space for preliminary programme and registration information, or sign up to the SIENNA newsletter to make sure  you receive the invitation.

Artificial Intelligence and Robotics:  Ethical, legal and human rights challenges.

In this part of the conference, we will discuss how ethics is currently being used and integrated in AI practices and institutions, diagnose problems and challenges, and will propose ways to move forward.  We will discuss new research ethics frameworks for AI, new regulatory and policy proposals, Ethics by Design methodologies, the role of stakeholders and the general public in responsible innovation for AI, the role of critical and social studies of AI in ethics and policy, and education and training programmes for ethics of AI.

Human Genomics:  Ethical, legal and human rights challenges

In this session, we will discuss the new ethical challenges brought by new human genomics technologies, particularly new technologies for genome sequencing and editing, and their various uses in diagnosis and therapy.  We will discuss regulatory challenges for human genomics, and possible solutions for them, and we will propose new instruments for ethical guidance and governance of genomic technologies, including a new SIENNA-initiated code of conduct for international data sharing in genomics.

Human Enhancement:  Ethical, legal and human rights challenges

In this session, we will discuss the ethical challenges emerging human enhancement technologies, and assess the current state of the field and the steps that should be taken to meet ethical challenges in the future.  We will present and discuss the new SIENNA-initiated ethics guidelines for human enhancement, the first extended set of guidelines for human enhancement research, development and use that have been proposed, and we will discuss needed new policy initiatives that are needed for responsible innovation of human enhancement technologies.

Assessment and Guidance of new emerging technologies

In this final session, we will propose new approaches for ethical and human rights assessment, guidance and governance of emerging technologies.  What steps do different actors need to take to enable responsible innovation, and what tools and methods do they need?  We will present and discuss proposals for integrating ethical guidance with regulation and policy, including stakeholders in ethical analysis and governance, for combining ethical analysis with foresight and social impact assessment, for the development of general and domain- and actor-specific ethics guidelines, standards and certification, and for including Ethics by Design methods in technology development.

Public consultation on ethical guidance for genomics, human enhancement, AI & robotics

SIENNA Public consultation on ethical guidance for genomics, human enhancement, AI & robotics 11-25 January

SIENNA invites you to participate in a public consultation on our proposals! Between 11-25 January we invite you to a public consultation of a group of documents with concrete ethical guidance for human genetics and genomics, human enhancement, artificial intelligence and robotics! Find out more about the process here, or go straight to the proposals you are interested in (listed below).

Human genomics

  • Operational guidance for ethical self-assessment of research

Human Enhancement

  • Ethical guidelines for research, development & application of human enhancement and proposals for the creation of an appropriate body to oversee and analyse trends towards human enhancement, assess moral and social consequences, provide information and advice.

Artificial intelligence & robotics

  • Ethics by design & ethics of use approaches for AI, robotics & big data
  • Industry education and buy-in
  • AI ethics education, training and awareness raising
  • Ethics as attention to context: recommendations for AI ethics

The documents that are now made public for feedback include:

The PIR sale

UPDATE: We won! After being warned by the California Attorney General ICANN could be in danger of violating its constitution by letting the sale go through, ICANN denied permission for the sale.

If you don’t know the background to the PIR sale and why it’s a BIG THING, here are a few links to fill you in.

Four US senators sent a letter to the Internet Society raising a number of questions regarding the sale of the Public Interest Registry to Ethos Capital.  Their letter can be found here: https://www.wyden.senate.gov/news/press-releases/wyden-blumenthal-warren-and-eshoo-question-sale-of-org-domains-to-private-equity-firm

Here’s ISOC’s response: https://www.keypointsabout.org/blog/the-internet-society-pir-and-ethos-respond-to-congressional-letter

Other background can be seen here:

Sam Klein published an article that summarises well many of the unanswered questions and concerns that ISOC members have about the sale of PIR:http://blogs.harvard.edu/sj/2019/12/02/the-dot-org-fire-sale-sold-for-half-its-valuation/

It’s a PRIVATE vs PUBLIC debate regarding internet governance

I think there are two different perspectives in conflict over this issue.

Perspective 1: the PIR is a business with customers

Perspective 2: the PIR is a public body, like a public utility, with users, not customers. 

I’ve spent much time in both Europe and the USA, and I think the two regions differ strongly over public services being provided by commercial companies.  This is best seen in their attitudes to public health.  Most Europeans regard the US provision of health care via profit-oriented businesses as almost insane, morally questionable (or even just plain evil) and blatantly inefficient; while most US citizens see nothing wrong with it, and either wonder how any government can possibly operate an effective health care system (because, in their view, only the profit motive makes for efficiency) or have a reflexive aversion to anything labelled “socialist.”

The role of the PIR as providing services for the non-profit sector further reinforces the perceptions of the “public service” people because non-profit is largely public service oriented.  For them, the PIR is not a business,  but a public body.  It’s like calling NASA a business.  For the public service perspective, this is a huge change.  So for these people, the appropriate model for evaluating the sale is the type of deal a government would do in awarding a contract for delivery of public services – generation of support from the stakeholders by explaining the case, consultation on solutions, public tenders with appropriate transparency, formal mechanisms for challenging the decision etc. 

From my experience of US culture, there is much less value given to government provision of many services, and much more faith that commercial businesses will act ethically, than I see in Europe.

This is why we see continual arguments over the process.  Those supporting the sale use the model of business negotiations, while those opposing it use the model of government tenders.

From the “public service perspective”, the primary responsibility of the PIR is to meet the needs of the user, and making profit is important only to the degree it helps meet those needs.  From this perspective, if it becomes profit-oriented, the PIR’s responsibility has to be that of deriving more value from registrants than it delivers, in order to pass that profit to shareholders.   We saw many arguments about this point.  The defence has been that the PIR will have to meet the registrant’s needs in order to make a profit.  However, the concern is that we have to trust it to do so, we can’t force it.  Many, especially in Europe, simply don’t trust any business that much – because it’s a business.  Furthermore, moving to a for-profit model introduces this potential conflict within PIR between service level and profitability.  Whether that conflict ever reaches the point of service decline is impossible to predict.  However, the risk doesn’t exist at all if the PIR is non-profit.  So the sale creates a new risk.  Many don’t think it’s acceptable for ISOC to create such risks. 

Ethos has made public moves to address this concern with talk of B-Corp certification and the Stewardship Council.  However, they have not made specific binding commitments.  They have made statements, but there is nothing legally holding them to any course of action.  Furthermore, Andrew has specifically stated there are no undisclosed terms in contracts binding Ethos, and that what Ethos does once the sale is complete will no longer be an area of ISOC responsibility.  Ethos haven’t actually formally committed to going B-Corp, but it wouldn’t actually address the issue anyway because B-Corp or B-Cert doesn’t stop you profit-gouging, reducing services etc.  So B-anything won’t legally bind to the degree necessary to address the fears.  For many, the Stewardship Council has not adequately addressed this issue either.  Firstly, Ethos appoint the initial members of the council, who then appoint themselves forever.  There’s no transparency and no representation of the stakeholders.  Secondly, Ethos have not specified what formal control the council will have over PIR decisions.  As far as I understand from their webinar appearances, it will be largely advisory.  But they have specifically stated the Stewardship Council will not have veto over price increases.  So the only concrete commitment Ethos have made about the council is that it won’t have the power to prevent the thing they are offering it as a solution to.

From public service perspective, arguments about giving the PIR the ability to expand services, increase customers etc, are irrelevant to assessing the deal.  To the public service view, not only do such arguments fail to justify the sale, making these arguments at all demonstrates an incorrect understanding of what the PIR is – a public service.  Making these arguments as a response to the public service perspective makes matters worse because it shows a failing to understand why the deal is considered objectionable by the public service group in the first place.  Pointing out .orgs have been managed by for-profits in the past simply repeats the same error – it’s like saying “things were bad before so it’s ok if we make them all over bad again.”

From the public service perspective, this is also a significant change in the overall structure of the internet.  Domain names are the root of the whole tree.  Domain names are the gateway to presenting yourself to the internet.  The ability to add or remove an organisation from the web is a significant power in a global society.  The effects of how and what is done within domain registration have subtle but huge repercussions throughout the web.  From this perspective, there is a structural need for some part of domain name registration to operate in a non-commercial fashion.  The reasons why are too many for here, but can be summarised as “a healthy society is impossible if all services are provided by commercial organisations.”

If you have a distrust of commercial provision of public services, multi-level shell corporations, expansive NDA’s and hiding director identities, all look highly suspicious.  It might be business as usual, in business.  But in a public service, expectations are different and so these business activities are themselves proof that the deal is inappropriate.

As I see it, the only way the public service perspective could be satisfied would be if there were legally binding limits on what Ethos could do.  These would have to cover selling the PIR to another, probably something about who could be a shareholder, powers of the board, etc, and .org price changes.  These commitments would have to be far more than most businesses would agree to and Ethos have shown every sign of avoiding such commitments.  

Those behind and supporting the sale, I suggest, do not see things this way.  For them the PIR is a business, making a profit is fine, it’s no big deal if every registration authority is a commercial enterprise.  For them, this is not a significant structural change in the internet.  I suggest this type of thinking is fairly common in the USA, but rare throughout most of Europe. I suggest one or the other perspective would dominate in most countries.

As a global organisation, I don’t think ISOC should pick a side between these two perspectives.  I think ISOC leadership needs to understand both perspectives, and I think it needs to give space to both.  People in ISOC governance need to recognise some members are ardent Marxists who would introduce a socialist state if they could, yet would be considered conservative amongst their even more Marxist community.  Other members would like to limit the role of profit-making companies to very limited sectors of the economy.  Other members think Marxism is as close to evil as a political ideology can be.  Some members would always trust the government to deliver a better service than any commercial company, while others assume every government enterprise must be blotted and inefficient.  ISOC has to accommodate all these extremes.  I do not see signs that it has done so.  My reading of the first emails on this issue is that ISOC leadership were surprised by the objections of the public service perspective, which indicates they had not realised this way of thinking existed within ISOC.  The arguments of about describing the PIR as a business, denying this was a significant change to overall internet governance, etc, suggest a way of thinking so deeply immersed in the commercial perspective that the public service perspective is almost incomprehensible.  I am not suggesting they should have agreed with it, merely they should have anticipated it, should understand it, and should be able to engage with it rather than simply asserting the opposite.  There’s no point having a global mission to connect everyone if we’re not going to use that as an opportunity to understand each other better.  And as a global organisation, ISOC needs to accommodate the widest range of perspectives possible.