ALT Winter Summit on Ethics and Artificial Intelligence

Last week I joined the ALT Winter Summit on Ethics and an Artificial Intelligence. Earlier in the year I was following developments at the interface between ethics, AI and the commons, which resulted in this blog post: Generative AI: Ethics all the way down.  Since then, I’ve been tied up with other things, so I appreciated the opportunity to turn my attention back to these thorny issues.  Chaired by Natalie Lafferty, University of Dundee, and Sharon Flynn, Technological Higher Education Association, both of whom have been instrumental in developing ALT’s influential Framework for Ethical Learning Technology, the online summit presented a wide range of perspectives on ethics and AI, both practical and philosophical, from scholars, learning technologists and students.  

Whose Ethics? Whose AI? A relational approach to the challenge of ethical AI – Helen Beetham

Helen Beetham opened the summit with an inspiring and thought-provoking keynote that presented the case for relational ethics. Positionality is important in relational ethics; ethics must come from a position, from somewhere. We need to understand how our ethics are interwoven with relationships and technologies. The ethics of AI companies come from nowhere. Questions of positionality and power engender the question “whose artificial intelligence”?  There is no definition of AI that does not define what intelligence is. Every definition is an abstraction made from an engineering perspective, while neglecting other aspects of human intelligence.  Some kinds of intelligence are rendered as important, as mattering, others are not. AI has always been about global power and categorising people in certain ways.  What are the implications of AI for those that fall into the wrong categories?

Helen pointed out that DARPA have funded AI intensively since the 1960’s, reminding me of many learning technology standards that have their roots in defence and aeronautical industries.

A huge amount of human refinement is required to produce training data models; this is the black box of human labour, mostly involving labourers in the global south.  Many students are also working inside the data engine in the data labelling industry. We don’t want to think about these people because it affects the magic of AI.

At the same time, tools are being offered to students to enable them to bypass AI detection, to ‘humanise” the output of AI tools.  The “sell” is productivity, that this will save students’ time, but who benefits from this productivity?

Helen noted that the terms “generative”, “intelligence”, and “artificial” are all very problematic and said she preferred the term “synthetic media”.  She argued that it’s unhelpful to talk about the skills humans need to work alongside AI, as these tools have no agency, they are not co-workers. These approaches create new divisions of labour among people, and new divisions about whose intelligence matters. We need a better critique of AI literacy and to think about how we can ask questions alongside our students. 

Helen called for universities to share their research and experience of AI openly, rather than building their own walled gardens, as this is just another source of inequity.  As educators we hold a key ethical space.  We have the ingenuity to build better relationships with this new technology, to create ecosystems of agency and care, and empower and support each other as colleagues.

Helen ended by calling for spaces of principled refusal within education. In the learning of any discipline there may need to be spaces of principled refusal, this is a privilege that education institutions can offer. 

Developing resilience in an ever-changing AI landscape ~ Mary Jacob, Aberystwyth University

Mary explored the idea of resilience and why we need it. In the age of AI we need to be flexible and adaptable, we need an agile response to emerging situations, critical thinking, emotional regulation, and we need to support and care for ourselves and others. AI is already embedded everywhere, we have little control over it, so it’s crucial we keep the human element to the forefront.  Mary urged us to notice our emotions and think critically, bring kindness and compassion into play, and be our real, authentic selves.  We must acknowledge we are all different, but can find common ground for kindness and compassion.  We need tolerance for uncertainty and imperfection and a place of resilience and strength.

Mary introduced Aberystwyth’s AI Guidance for staff and students and also provided a useful summary of what constitutes AI literacy at this point in time.

Mary Jacob's AI Literacy

Achieving Inclusive education using AI – Olatunde Duruwoju, Liverpool Business School

Tunde asked us how we address gaps in inequity and inclusion?  Time and workload are often cited as barriers that prevent these issues from being addresses, however AI can help reduce these burdens by improving workflows and capacity, which in turn should help enable us to achieve inclusion.

When developing AI strategy, it’s important to understand and respond to your context. That means gathering intersectional demographic data that goes beyond protected characteristics.  The key is to identify and address individual students issues, rather than just treating everyone the same. Try to understand the experience of students with different characteristics.  Know where your students are coming from and understand their challenges and risks, this is fundamental to addressing inclusion.

AI can be used in the curriculum to achieve inclusion.  E.g. Using AI can be helpful for international students who may not be familiar with specific forms of assessment. Exams trigger anxiety, so how do we use AI to move away from exams?

Olatunde Duruwoju - Think intersectionality

AI Integration & Ethical Reflection in Teaching – Tarsem Singh Cooner

Tarsem presented a fascinating case study on developing a classroom exercise for social work students on using AI in practice.  The exercise drew on the Ethics Guidelines on Reliable AI from the European Group on Ethics, Science and New Technologies and mapped this against the Global Social Work Ethical Principles.

Tarsem Singh Cooner - comparison of Principles on Reliable AI  and Global Social Work Ethical Principles

The assignment was prompted by the fact that practitioners are using AI to uncritically write social work assessments and reports. Should algorithms be used to predict risk and harm, given they encode race and class bias? The data going into the machine is not benign and students need to be aware of this.

GenAI and the student experience – Sue Beckingham, Louise Drum, Peter Hartley & students

Louise highlighted the lack to student participation in discussions around AI. Napier University set up an anonymous padlet to allow students to tell them what they thought. Most students are enthusiastic about AI. They use it as a dialogue partner to get rapid feedback. It’s also helpful for disabled and neurodivergent students, and those who speak English as a second language, who use AI as an assistive technology.  However students also said that using AI is unfair and feels like cheating.  Some added that they like the process of writing and don’t want to loose that, which prompted Louise to ask if we’re outsourcing the process of critical thinking?  Louise encouraged us to share our practice through networks, adding that collaboration and cooperation is key and can lead to all kinds of serendipity.

The students provided a range of different perspectives:

Some reported conflicting feelings and messages from staff about whether and how AI can be used, or whether it’s cheating.  Students said they felt they are not being taught how to use AI effectively.

GCSEs and the school system just doesn’t work for many students, not just neurotypical ones, it’s all about memorising things.  We need more skills based learning rather than outcome based learning.

Use of AI tools echoes previous concerns about the use of the internet in education. There was a time when there was considerable debate about whether the internet should be used for teaching & learning.

AI can be used to support new learning. It provides on hand personal assistance that’s there 24/7.  Students create fictional classmates and partners who they can debate with.  A lot of it is garbage but some of it is useful. Even when it doesn’t make sense, it makes you think about other things that do make sense.

A few thoughts…

As is often the case with any new technology, many of the problematic issues that AI has thrown up relate less to the technology itself, and more to the nature of our educational institutions and systems.  This is particularly true in the cases of issues relating to equity, diversity and inclusion; whose knowledge and experiences are valued, and whose are marginalised?   

It’s notable that several speakers mentioned the use of AI in recruitment. Sue Beckingham noted that AI can be helpful for interview practice, though Helen highlighted research that suggested applicants who used chatGPT’s paid functionality perform much better in recruitment than those who don’t.  This suggests that we need to be thinking about authentic recruitment practices in much the same way we think about authentic assessment.  Can we create recruitment process that mitigate or bypass the impact of these systems?

I particularly liked Helen’s characterisation of AI as synthetic media, which helps to defuse some of the hype and sensationalism around these technologies.

The key to addressing many of the issues relating to the use of AI in education is to share our practice and experience openly and to engage our colleagues and students in conversations that are underpinned by contextual ethical frameworks such as ALT’s Framework for Ethical Learning Technology.  Peter Hartley noted that universities that have already invested in student engagement and co-creation are at an advantage when it comes to engaging with AI tools.

I’m strongly in favour of Helen’s call for spaces of principled refusal, however at the same time we need to be aware that the genie is out of the bottle.  These tools are out in the world now, they are in our education institutions, and they are being used by students in increasingly diverse and creative ways, often to mitigate the impact of systemic inequities. While it’s important to acknowledge the exploitative nature and very real harms perpetrated by the AI industry, the issues and potential raised by these tools also give us an opportunity to question and address systemic inequities within the academy. AI tools provide a valuable starting point to open conversations about difficult ethical questions about knowledge, understanding and what it means to learn and be human.  

Generative AI – Ethics all the way down

How to respond to the affordances and challenges of generative AI is a pressing issue that many learning technologists and open education practitioners are grappling with right now and I’ve been wanting to write a blog post about the interface between AI, large language models and the Commons for some time. This isn’t that post.  I’ve been so caught up with other work that I’ve barely scratched the surface of the articles on my rapidly expanding reading list.  Instead, these are some short, sketchy notes about the different ethical layers that we need to consider when engaging with AI.  This post is partly inspired by technology ethics educator Casey Fiesler, who has warned education institutions of the risk of what she refers to as ethical debt. 

“What’s accruing here is not just technical debt, but ethical debt. Just as technical debt can result from limited testing during the development process, ethical debt results from not considering possible negative consequences or societal harms. And with ethical debt in particular, the people who incur it are rarely the people who pay for it in the end.”
~ Casey Fiesler, The Conversation

Apologies for glossing over the complexity of these issues, I just wanted to get something down in writing while it’s fresh in my mind 

Ethics of large language models and Common Crawl data sets

Most generative AI tools use data sets scraped from the web and made available for research and commercial development.  Some of the organisations creating these data sets are non-profits, others are commercial companies, the relationship between the two is not always transparent. Most of these data sets scrape content directly from the web regardless of ownership, copyright, licensing and consent, which has led to legitimate concerns about all kinds of rights violations. While some companies claim to employ these data sets under the terms of fair use, questions have been raised about using such data for explicitly commercial purposes. Some open advocates have said that while they have no objection to these data sets being used for research purposes they are very concerned about commercial use. Content creators have also raised objections to their creative works being used to train commercial applications without their knowledge or consent.  As a result, a number copyright violation lawsuits have been raised by artists, creators, cultural heritage organisations and copyright holders.

There are more specific issues relating to these data sets and Creative Commons licensed content.  All CC licenses include an attribution clause, and in order to use a CC licensed work you must attribute the creator. LLMs and other large data sets are unable to fulfil this crucial attribution requirement so they ride roughshod over one of the foundational principles of Creative Commons. 

LLMs and common crawl data sets are out there in the world now.  The genie is very much out of the bottle and there’s not a great deal we can do to put it back, even if we wanted to. It’s also debatable what, if anything, content creators, organisations and archives can do to prevent their works being swept up by web scraping in the future. 

Ethics of content moderation and data filtering

Because these data sets are scraped wholesale from the web, they inevitably include all kinds of offensive, degrading and discriminatory content. In order to ensure that this content does not influence the outputs of generative AI tools and damage their commercial potential, these data sets must be filtered and moderated.  Because AI tools are not smart enough to filter out this content automatically, the majority of content moderation is done by humans, often from the global majority, working under exploitative and extractive conditions. In May, content moderators in Africa who provide services for Meta, Open AI and others voted to establish the first African Content Moderators Union, to challenge low pay and exploitative working conditions in the industry. 

Most UK universities have a commitment to ending modern slavery and uphold the terms of the Modern Slavery Act. For example the University of Edinburgh’s Modern Slavery Statement says that it is “committed to protecting and respecting human rights and have a zero-tolerance approach to slavery and human trafficking in all its forms.” It is unclear how commitments such as these relate to content workers who often work under conditions that are exploitative and degrading at best, and a form of modern slavery at worst. 

Ethics of anthropomorphising AI 

The language used to describe generative AI tools often humanises and anthropomorphises them, either deliberately or subconsciously. They are ascribed human characteristics and abilities, such as intelligence and the ability to dream. One of the most striking examples is the use of hallucinating.  When Chat GPT makes up non-existent references to back up erroneous “facts” this is often described as “hallucinating“.  This propensity has led to confusion among some users when they have attempted to find these fictional references. Many commenters have pointed out that these tools are incapable of hallucinating, they’re just getting shit wrong, and that the use of such humanising language purposefully disguises and obfuscates the limitations of these systems. 

“Hallucinate is the term that architects and boosters of generative AI have settled on to characterize responses served up by chatbots that are wholly manufactured, or flat-out wrong.”
~ Naomi Klein, The Guardian

Ethics of algorithmic bias

Algorithmic bias is a well known and well documented phenomenon (cf Safiya U. Noble‘s Algorithms of Oppression) and generative AI tools are far from immune to bias. Valid arguments have been made about the bias of the ‘intelligence” these tools claim to generate.  Because the majority of AI applications are produced in the global north, they invariably replicate a particularly white, male, Western world view, with all the inherent biases that entails. Diverse they are not. Wayne Holmes has noted that AI ignores minority opinions and marginalised perspectives, perpetuating a Silicon Valley perspective and world outlook. Clearly there are considerable ethical issues about education institutions that have a mission to be diverse and inclusive using tools that engender harmful biases and replicate real world inequalities. 

“I don’t want to say I’m sure. I’m sure it will lift up the standard of living for everybody, and, honestly, if the choice is lift up the standard of living for everybody but keep inequality, I would still take that.”
~ Sam Altman, OpenAI CEO. 

Ethics of catastrophising
 

Much has been written about the dangers of AI, often by the very individuals who are responsible for creating these tools. Some claim that generative AI will end education as we know it, while others prophesy that AI will end humanity altogether. There is no doubt that this catastrophising helps to feed the hype cycle and drive traffic to to these tools and applications, however Timnit Gebru and others have pointed out that by focusing attention on some nebulous future catastrophe, the founding fathers of AI are purposeful distracting us from current real world harms caused by the industry they have created, including reproducing systems of oppression, worker exploitation, and massive data theft. 

“The harms from so-called AI are real and present and follow from the acts of people and corporations deploying automated systems. Regulatory efforts should focus on transparency, accountability and preventing exploitative labor practices.”
Nirit Weiss-Blatt’s (@DrTechlash) “Taxonomy of AI Panic Facilitators” A visualization of leading AI Doomers (X-risk open letters, media interviews & OpEds). Some AI experts enable them, while others oppose them. The gender dynamics are fucked up. It says a lot about the panic itself.

Not really a conclusion

Clearly there are many ethical issues that education institutions must take into consideration if they are to use generative AI tools in ways that are not harmful.  However this doesn’t mean that there is no place for AI in education, far from it.  Many AI tools are already being used in education, often with beneficial results, captioning systems are just one example that springs to mind.  I also think that generative AI can potentially be used as an exemplar to teach complex and nuanced issues relating to the creation and consumption of information, knowledge equity, the nature of creativity, and post-humanism.  Whether this potential outweighs the ethical issues remains to be seen.

A few references 

AI has social consequences, but who pays the price? Tech companies’ problem with ‘ethical debt’ ~ Casey Fiesler, The Conversation 

Statement from the listed authors of Stochastic Parrots on the “AI pause” letter ~ Timnit Gebru (DAIR), Emily M. Bender (University of Washington), Angelina McMillan-Major (University of Washington), Margaret Mitchell (Hugging Face)

Open letter to News Media and Policy Makers re: Tech Experts from the Global Majority ~ @safiyanoble (Algorithms of Oppression), @timnitGebru (ex Ethical Artificial Intelligence Team)@dalitdiva, @nighatdad, @arzugeybulla, @Nanjala1, @joana_varon

150 African Workers for ChatGPT, TikTok and Facebook Vote to Unionize at Landmark Nairobi Meeting ~ Time

AI machines aren’t ‘hallucinating’. But their makers are ~ Naomi Klein, The Guardian  

Just Because ChatBots Can’t Think Doesn’t Mean They Can’t Lie ~ Maria Bustillos, The Nation 

Artificial Intelligence and Open Education: A Critical Studies Approach ~ Dr Wayne Holmes, UCL 

‘What should the limits be?’ The father of ChatGPT on whether AI will save humanity – or destroy it ~ Sam Altman interview, The Guardian