Ethics & Policy

Analysing AI Ethics… Using AI!

Published

6 days ago

July 1, 2025

peifers

By: Alex Cline, Alice Helliwell, Brian Ball, David Freeborn and Kevin Loi-Heng

Project supported by the Ethics Institute, the Internet Democracy Initiative and NULab, at Northeastern University

What can computational methods–particularly artificial intelligence (AI)–tell us about AI ethics? We are a group of researchers who have become intrigued by what such methods can add to philosophical studies, and vice versa. In the programmes at Northeastern University London,^[1] we are interested in AI ethics. Education in this subject forms an integral part of our teaching, particularly on our MA Philosophy and Artificial Intelligence and our MSc AI and Ethics. Our students have a wide range of backgrounds, (though many have been trained in either philosophy or computer/data science). Our aim is to provide philosophical and computational education simultaneously, to equip students with the skills they need to responsibly engage with AI technology.

Given this ethos, we have decided to turn use of computational methodologies on our own practice, by investigating some of our philosophy courses on these programmes. Our aim is to gain insight into our pedagogical approach and to develop a project which we can (hopefully) share with our students. In fact, one of the researchers on this project (Kevin Loi-Heng) is an alumnus of our MSc! So far, we have found this process to be surprising and rewarding.

In order to test our thought that computational tools can be useful for pedagogical and philosophical goals, we decided to conduct a computational analysis of the texts we set for students across two courses: AI and Data Ethics, and Advanced Topics in Responsible AI. We have (rather grandiosely) called this our ‘canon analysis’. Of course, we did not gather these papers with the intention that they truly be a ‘canon’ of AI ethics. We have curated them over several years, and after completing both courses, we want our students to have covered a variety of classic and current topics in AI ethics and responsible AI. Having gathered the recommended texts for these courses, we utilised some standard Python-based natural language processing (NLP) techniques to analyse our corpus of texts.

Absolute Word-frequency Analysis

The first analytic tool we turned on our corpus was a word frequency counter. This simple computational technique counts the number of times a word appears in a document, or collection of documents. This allowed us to identify the words that appear most frequently in our collection of papers, and produce the word cloud below (where the most frequently used words appear largest in size):

Image 1: word cloud of frequently appearing words in the corpus (aka ‘the canon’).

To our surprise, we found that across our two courses, the highest frequency unique word used was ‘human’! Perhaps, as scholars in the ‘humanities’ this shouldn’t have been unexpected to us, however we consider this corpus of texts to be primarily concerned with technology and ethics. It may be that the authors used on our courses are contrasting humans with the data (2nd most common) and machines (8th most common) that are their explicit focuses.

Term frequency itself has limited utility for telling us about unique features of a corpus of texts. It could be, for example, that (contrary to the conjecture at the end of the last paragraph) ‘human’ is something that comes up in philosophical works in general. To find out more about the unique features of this body of texts, we conducted another analysis.

Relative Word Frequency Analysis

Surprised at the most frequent word being ‘human’ we decided to analyse word-frequency further. We ran another measure on the corpus: a TF-IDF (Term Frequency–Inverse Document Frequency). This NLP technique is typically used to evaluate the relative importance of a word in a document compared to its importance in the corpus as a whole. Rather than simply counting the frequency of use for each word, a TF-IDF can show which words are more common in our AI Ethics corpus compared to a larger, or alternative, corpus of texts.

Of course, in this case we were not just interested here in individual papers, but the body of works (our AI ethics ‘canon’) as a whole. In order to complete a TF-IDF measure then, we required a contrasting corpus of texts. It just so happened that, following the Wittgenstein and AI conference and edited collections some of us had recently produced (see volume 1 and volume 2), we had a ‘Wittgenstein Corpus’ available; a body of papers (accessed through JSTOR) discussing the work of Wittgenstein.

Top 10 words: AI Ethics Canon	Top 10 words: Wittgenstein Corpus
human	philosophy
ethic	Wittgenstein
moral	philosophical
robot	language
data	theory
system	political
technology	social
design	review
agent	science
develop	knowledge

Image 2: Top 10 Words for AI Ethics Canon and Wittgenstein Corpus

When we compare these two analyses, we start to see the relative importance of these terms in the text. ‘Human’, for example, is not just the most frequent unique word, but it’s particularly important in the AI ethics papers compared to works discussing Wittgenstein. ‘Wittgenstein’ is the second most important word in the Wittenstein papers (a comforting sign that our analysis was working).

Using AI: Semantic clustering

Few nowadays would consider the techniques we have discussed so far to involve AI: in particular, the computational methods employed operate directly on textual data, here the full papers from our two course reading lists. Since research papers are written in natural language, they need to be converted into a numerical format that a computer can read and interpret if contemporary AI techniques are to be deployed on them. We did this using SciBERT, a state-of-the-art transformer model pre-trained on scientific texts. This allowed us to turn each paper into a unique vector (essentially a high-dimensional mathematical fingerprint) that captures the meaning of the text.

Once we had numerical representations of each paper, we compared them to each other using cosine similarity. This helped us measure how closely related different papers are based on their content. A similarity score of 1 means two papers are essentially the same, while a score close to 0 means they are very different.

Using this measure of similarity we attempted to draw out where papers in our canon were grouped together around different subjects and themes. To examine this, we utilised a couple of methods. First, we applied K-Means clustering ,an unsupervised machine learning technique that groups papers into clusters based on their similarity. It works on unlabeled data (without defined categories or groups). The algorithm first randomly selects central points (centroids), then uses algorithms to automatically find common themes and structures in the data. We repeated the clustering with different k values to find different groupings. By experimenting with different k values we determined the best number of clusters. For this we used techniques like the Elbow Method and Silhouette Score to find a suitable number given the tradeoff between better representing the data and using more clusters. We decided on six clusters to move forwards.

Image 3: Silhouette Score vs. Number of Clusters

K-means struggles to deal with as many dimensions as provided by SciBERT’s analysis of the papers, so we had to process the data further. We did a principal component analysis (PCA) to reduce the dimensionality of the data. PCA is a linear algebra technique which finds directions in the data that can explain the greatest proportion of variance. We utilised 112 components, as this explained 95% of the variance in our data. Having reduced the dimensionality of our data, we conducted K-means clustering.

Image 4: Explained variance vs. Principal components.

To ensure that the clustering results were meaningful, we checked whether each paper had the highest similarity to the average of its assigned cluster. The fact that 100% of papers were most similar to their own cluster’s average reassured us that the model was making reasonable groupings. In order to visualise these clusters, we needed to conduct further processing on this data, again using PCA, to reduce the clusters to two dimensions. This two dimensional data could then be visualised on a series of graphs, using difference

Image 5: K-means clustering with principal component analysis showing the division of papers depending on different numbers of clusters. The dots represent the papers in the canon, with colours representing the clusters to which they belong.

When we looked at which papers fell in each cluster, however, we had a hard time interpreting these clusters. We couldn’t clearly determine which topic/s in AI ethics were key for each cluster. This was likely due to the high dimensionality, and the small number of papers included in our analysis. We are reminded that contemporary AI relies on BIG data! We therefore tried an alternative method for grouping the papers in our canon.

Using AI: Topic Analysis

We next used the Latent Dirichlet Allocation (LDA) method to examine the canon, to see if the paper groupings produced made more sense to us. LDA is also an unsupervised machine learning approach. However, unlike K-means, we can use LDA to gather papers under topics, and to then produce a list of words for each topic, making it more interpretable.

LDA is a soft clustering method, which models probability distributions over words and documents. When we use LDA to analyse papers, it treats each paper as a collection of words – i.e. it does not consider the position of each word in the paper (unlike SciBERT). LDA builds a model of the whole corpus, and tries to identify distinct topics by finding correlations between words. Frequent co-occurrence of words suggests they are related in a topic, whereas non-co-occurrence of words suggests they are not related in a topic.

Our output from LDA is a series of probabilities. For each paper we get a probability that it falls in each topic (six topics). A paper is therefore not just assigned to one topic – instead, it can have a high probability of concerning multiple topics. This may be for good reason – for example, an overview paper might end up having a high probability of concerning e.g. ‘privacy’ ‘AI design’ and ‘robot agency’ (etc).

From examining the topics uncovered in this manner, we felt like we could make some sense of them. We identified the broad themes of each topic as follows:

Topic clusters:

0: Social, social media, gender, culture

1: Superintelligence

2: Applied issues such as sustainability, health, and the arts

3: Robots, personhood, and artificial agency

4: Design, responsibility

5: Privacy and risk

Image 6: Topic Overlap Matrix (Number of Papers in Common)

These topics certainly seemed to us to have some internal unity (as indicated), but they could also be seen not to overlap one another in problematic ways. Looking at the percentage of the papers in one topic (the row in the above table) that overlapped with papers in the other topic (in the columns), we found both that the overlap was not in general too great, and that such overlap as there was could be readily interpreted. For example, 52.9% of the papers on superintelligence could also be viewed as concerned with a topic involving the notion of artificial agency, which is understandable given that ethical concerns around the former appeal to the latter; moreover, looking at the column corresponding to superintelligence, we see that it is entirely blue, meaning that none of the other topics overlapped much with it – and indeed, our impression from working within the field is that this topic does, as a matter of sociological fact about the AI ethics community, stand somewhat apart.

What can AI tell us about philosophy

AI-powered knowledge discovery is being widely applied in STEM fields like biology and genomics, and for drug-discovery etc.; but these advancements have not, on the whole, extended into the humanities and social sciences, including philosophy (which has historically played an integral part in AI development). Some progress has been made in using NLP techniques, with the availability of pre-trained LLMs which offer some promising utility to help process textual data. Our research leverages these methods to begin to make sense of a growing academic literature in AI ethics – at least how we have presented it to students. In the future, we hope to continue our analysis of AI ethics literature, and share this with our students to gain their perspective of how this analysis meshes with their understanding of our courses.

^[1] Alex Cline has now started working for Queen Mary University London.

Source link

Ethics & Policy

AI and ethics – what is originality? Maybe we’re just not that special when it comes to creativity?

Published

3 hours ago

July 7, 2025

Stuart Lauchlan

I don’t trust AI, but I use it all the time.

Let’s face it, that’s a sentiment that many of us can buy into if we’re honest about it. It comes from Paul Mallaghan, Head of Creative Strategy at We Are Tilt, a creative transformation content and campaign agency whose clients include the likes of Diageo, KPMG and Barclays.

Taking part in a panel debate on AI ethics at the recent Evolve conference in Brighton, UK, he made another highly pertinent point when he said of people in general:

We know that we are quite susceptible to confident bullshitters. Basically, that is what Chat GPT [is] right now. There’s something reminds me of the illusory truth effect, where if you hear something a few times, or you say it here it said confidently, then you are much more likely to believe it, regardless of the source. I might refer to a certain President who uses that technique fairly regularly, but I think we’re so susceptible to that that we are quite vulnerable.

And, yes, it’s you he’s talking about:

I mean all of us, no matter how intelligent we think we are or how smart over the machines we think we are. When I think about trust, – and I’m coming at this very much from the perspective of someone who runs a creative agency – we’re not involved in building a Large Language Model (LLM); we’re involved in using it, understanding it, and thinking about what the implications if we get this wrong. What does it mean to be creative in the world of LLMs?

Genuine

Being genuine, is vital, he argues, and being human – where does Human Intelligence come into the picture, particularly in relation to creativity. His argument:

There’s a certain parasitic quality to what’s being created. We make films, we’re designers, we’re creators, we’re all those sort of things in the company that I run. We have had to just face the fact that we’re using tools that have hoovered up the work of others and then regenerate it and spit it out. There is an ethical dilemma that we face every day when we use those tools.

His firm has come to the conclusion that it has to be responsible for imposing its own guidelines here to some degree, because there’s not a lot happening elsewhere:

To some extent, we are always ahead of regulation, because the nature of being creative is that you’re always going to be experimenting and trying things, and you want to see what the next big thing is. It’s actually very exciting. So that’s all cool, but we’ve realized that if we want to try and do this ethically, we have to establish some of our own ground rules, even if they’re really basic. Like, let’s try and not prompt with the name of an illustrator that we know, because that’s stealing their intellectual property, or the labor of their creative brains.

I’m not a regulatory expert by any means, but I can say that a lot of the clients we work with, to be fair to them, are also trying to get ahead of where I think we are probably at government level, and they’re creating their own frameworks, their own trust frameworks, to try and address some of these things. Everyone is starting to ask questions, and you don’t want to be the person that’s accidentally created a system where everything is then suable because of what you’ve made or what you’ve generated.

Originality

That’s not necessarily an easy ask, of course. What, for example, do we mean by originality? Mallaghan suggests:

Anyone who’s ever tried to create anything knows you’re trying to break patterns. You’re trying to find or re-mix or mash up something that hasn’t happened before. To some extent, that is a good thing that really we’re talking about pattern matching tools. So generally speaking, it’s used in every part of the creative process now. Most agencies, certainly the big ones, certainly anyone that’s working on a lot of marketing stuff, they’re using it to try and drive efficiencies and get incredible margins. They’re going to be on the race to the bottom.

But originality is hard to quantify. I think that actually it doesn’t happen as much as people think anyway, that originality. When you look at ChatGPT or any of these tools, there’s a lot of interesting new tools that are out there that purport to help you in the quest to come up with ideas, and they can be useful. Quite often, we’ll use them to sift out the crappy ideas, because if ChatGPT or an AI tool can come up with it, it’s probably something that’s happened before, something you probably don’t want to use.

More Human Intelligence is needed, it seems:

What I think any creative needs to understand now is you’re going to have to be extremely interesting, and you’re going to have to push even more humanity into what you do, or you’re going to be easily replaced by these tools that probably shouldn’t be doing all the fun stuff that we want to do. [In terms of ethical questions] there’s a bunch, including the copyright thing, but there’s partly just [questions] around purpose and fun. Like, why do we even do this stuff? Why do we do it? There’s a whole industry that exists for people with wonderful brains, and there’s lots of different types of industries [where you] see different types of brains. But why are we trying to do away with something that allows people to get up in the morning and have a reason to live? That is a big question.

My second ethical thing is, what do we do with the next generation who don’t learn craft and quality, and they don’t go through the same hurdles? They may find ways to use {AI] in ways that we can’t imagine, because that’s what young people do, and I have faith in that. But I also think, how are you going to learn the language that helps you interface with, say, a video model, and know what a camera does, and how to ask for the right things, how to tell a story, and what’s right? All that is an ethical issue, like we might be taking that away from an entire generation.

And there’s one last ‘tough love’ question to be posed:

What if we’re not special? Basically, what if all the patterns that are part of us aren’t that special? The only reason I bring that up is that I think that in every career, you associate your identity with what you do. Maybe we shouldn’t, maybe that’s a bad thing, but I know that creatives really associate with what they do. Their identity is tied up in what it is that they actually do, whether they’re an illustrator or whatever. It is a proper existential crisis to look at it and go, ‘Oh, the thing that I thought was special can be regurgitated pretty easily’…It’s a terrifying thing to stare into the Gorgon and look back at it and think,’Where are we going with this?’. By the way, I do think we’re special, but maybe we’re not as special as we think we are. A lot of these patterns can be matched.

My take

This was a candid worldview that raised a number of tough questions – and questions are often so much more interesting than answers, aren’t they? The subject of creativity and copyright has been handled at length on diginomica by Chris Middleton and I think Mallaghan’s comments pretty much chime with most of that.

I was particularly taken by the point about the impact on the younger generation of having at their fingertips AI tools that can ‘do everything, until they can’t’. I recall being horrified a good few years ago when doing a shift in a newsroom of a major tech title and noticing that the flow of copy had suddenly dried up. ‘Where are the stories?’, I shouted. Back came the reply, ‘Oh, the Internet’s gone down’. ‘Then pick up the phone and call people, find some stories,’ I snapped. A sad, baffled young face looked back at me and asked, ‘Who should we call?’. Now apart from suddenly feeling about 103, I was shaken by the fact that as soon as the umbilical cord of the Internet was cut, everyone was rendered helpless.

Take that idea and multiply it a billion-fold when it comes to AI dependency and the future looks scary. Human Intelligence matters

Source link

Ethics & Policy

Experts gather to discuss ethics, AI and the future of publishing

Published

12 hours ago

July 7, 2025

文一

Representatives of the founding members sign the memorandum of cooperation at the launch of the Association for International Publishing Education during the 3rd International Conference on Publishing Education in Beijing.CHINA DAILY

Publishing stands at a pivotal juncture, said Jeremy North, president of Global Book Business at Taylor & Francis Group, addressing delegates at the 3rd International Conference on Publishing Education in Beijing. Digital intelligence is fundamentally transforming the sector — and this revolution will inevitably create “AI winners and losers”.

True winners, he argued, will be those who embrace AI not as a replacement for human insight but as a tool that strengthens publishing’s core mission: connecting people through knowledge. The key is balance, North said, using AI to enhance creativity without diminishing human judgment or critical thinking.

This vision set the tone for the event where the Association for International Publishing Education was officially launched — the world’s first global alliance dedicated to advancing publishing education through international collaboration.

Unveiled at the conference cohosted by the Beijing Institute of Graphic Communication and the Publishers Association of China, the AIPE brings together nearly 50 member organizations with a mission to foster joint research, training, and innovation in publishing education.

Tian Zhongli, president of BIGC, stressed the need to anchor publishing education in ethics and humanistic values and reaffirmed BIGC’s commitment to building a global talent platform through AIPE.

BIGC will deepen academic-industry collaboration through AIPE to provide a premium platform for nurturing high-level, holistic, and internationally competent publishing talent, he added.

Zhang Xin, secretary of the CPC Committee at BIGC, emphasized that AIPE is expected to help globalize Chinese publishing scholarships, contribute new ideas to the industry, and cultivate a new generation of publishing professionals for the digital era.

Themed “Mutual Learning and Cooperation: New Ecology of International Publishing Education in the Digital Intelligence Era”, the conference also tackled a wide range of challenges and opportunities brought on by AI — from ethical concerns and content ownership to protecting human creativity and rethinking publishing values in higher education.

Wu Shulin, president of the Publishers Association of China, cautioned that while AI brings major opportunities, “we must not overlook the ethical and security problems it introduces”.

Catriona Stevenson, deputy CEO of the UK Publishers Association, echoed this sentiment. She highlighted how British publishers are adopting AI to amplify human creativity and productivity, while calling for global cooperation to protect intellectual property and combat AI tool infringement.

The conference aims to explore innovative pathways for the publishing industry and education reform, discuss emerging technological trends, advance higher education philosophies and talent development models, promote global academic exchange and collaboration, and empower knowledge production and dissemination through publishing education in the digital intelligence era.

yangyangs@chinadaily.com.cn

Source link

Ethics & Policy

Experts gather to discuss ethics, AI and the future of publishing

Published

13 hours ago

July 7, 2025

文一

True winners, he argued, will be those who embrace AI not as a replacement for human insight but as a tool that strengthens publishing”s core mission: connecting people through knowledge. The key is balance, North said, using AI to enhance creativity without diminishing human judgment or critical thinking.

Tian Zhongli, president of BIGC, stressed the need to anchor publishing education in ethics and humanistic values and reaffirmed BIGC’s commitment to building a global talent platform through AIPE.

BIGC will deepen academic-industry collaboration through AIPE to provide a premium platform for nurturing high-level, holistic, and internationally competent publishing talent, he added.

Wu Shulin, president of the Publishers Association of China, cautioned that while AI brings major opportunities, “we must not overlook the ethical and security problems it introduces”.

yangyangs@chinadaily.com.cn

Source link