This is what it produced when I asked it to summarise my blog. The responses are still relatively limited but it gives a sense of what this technology will be able to do in the near future:
arkCarrigan.net is the professional website of Mark Carrigan, a digital sociologist at the Manchester Institute for Education. He leads a master’s degree program in digital education and co-leads an associated research group. His research interests involve the social ontology of emerging technologies, emerging technologies in higher education, and social media policy in the sector. He has written several books on these subjects and is currently working on two book projects and a research project focused on the visibility of social media in impact case studies12.
The website contains several sections, including “About Me”, “Research”, “Projects”, “Publications”, “Media Coverage”, and “Public Writing”. The “Research” and “Projects” sections feature blog-style entries that provide insights into his ongoing research and project initiatives. Due to technical limitations, the exact content of these sections could not be determined in detail.
The “Publications” section presumably contains information about Carrigan’s published works, but the specific details were not visible due to technical limitations. The “Media Coverage” section lists recent interviews with different organizations, including New Books Network, Good Practice in Teaching Podcast, Chinese Social Sciences Today, and Inside Higher Education3.
The “Public Writing” section is a blog where Carrigan writes on a variety of topics. Some of the titles include “How can academics use ChatGPT to make their work easier?”, “Current mood in AI generated images #1”, “Fully automated luxury universities”, “Who is paying for Twitter blue?”, and “Why do people who worry about the existential risks of AGI refuse to talk about capitalism?” among others4.
I asked for a more detailed analysis and this is what it provided. I’m intrigued about the topics identified here. Is there a capacity for topic modelling here which is being constrained for resourcing reasons?
he blog on markcarrigan.net covers a wide range of topics related to sociology, digital education, and emerging technologies. A summary of the topics from a few posts include:
- Andrew Sayer, phronesis, practical reasoning, technological reflexivity1.
- Platforms, public sociology, Social Media2.
- Generative AI and the future of universities3.
- Commonplace book, Ontology of emerging technologies, Platform & Agency4.
Given the variety in these topics, it’s not immediately clear which is the most frequent. However, it’s evident that the topics generally revolve around technology (particularly emerging technologies and AI), sociology, and higher education. These themes align with Mark Carrigan’s research interests as described on the “About Me” page of his website, where he mentions his focus on the social ontology of emerging technologies, emerging technologies in higher education, and social media policy in the secto
I then asked it to summarise what academics are arguing about Chat-GPT on the LSE Impact Blog, citing this by name rather than URL:
Writers on the LSE Impact Blog are discussing the implications of large language models (LLMs) like ChatGPT in the context of the evolving knowledge economy. The main arguments and points of discussion are as follows:
- Rapid Innovation: The pace of innovation in LLMs like ChatGPT is rapid, and it’s challenging to predict how these advancements will impact future skills needs and the knowledge economy at large. The need for sophisticated moral judgement in managing these tools is highlighted12.
- Transformation of the Knowledge Economy: LLMs are expected to transform the knowledge economy, particularly the roles of ‘symbolic analysts’, who solve, identify, and broker problems by manipulating symbols. The same features that protected these jobs from routine automation may now expose them to a new generation of technology that’s increasingly adept at symbol manipulation. There are very few symbolic analyst outputs that LLMs won’t be able to draft3.
- Potential Job Impact: LLMs might influence around 80% of US workers, with the highest impact expected for higher-qualified and higher-paid workers. However, it’s noted that AI technologies like ChatGPT won’t necessarily ‘destroy jobs’ but are likely to create new roles and remodel others. There’s a particular skill required to manage these machines, interpret their outputs, and evaluate their responses45.
- Professional Development: The automation of menial tasks by LLMs may disrupt professional development, especially for new entrants in professions that involve preparing legal notes, due diligence reports, etc. If material is increasingly commissioned from machines, it could pose challenges not only for future graduates but also for those at the top of the professions who rely on a stream of trainees to replace them. It might affect the value of cognitive labor6.
- Erosion of Doubt: In the longer term, the issue may not be the inaccuracy of LLM outputs, but the erosion of doubt. While significant efforts have been made to prevent LLMs from generating inappropriate content, future models may be much clearer about the ‘right answer’ to any question. This can be helpful but the lack of transparency and contestability is also seen as a concern7.
- Ethics and AI: There’s a focus on the ethical, political, and legal dimensions of technology like ChatGPT. Minority groups and their representation in AI, and the concentration of AI power in industry and academia, are issues being discussed.
There was a specific discussion about minority reports and AI power concentration that I started to explore but wasn’t able to finish due to time constraints. I recommend looking into these specific areas further for a more comprehensive understanding of what writers on the LSE Impact Blog are arguing about ChatGPT.
Mark Carrigan 