Can we trust AI with research?

Artificial intelligence is still the talk of the town. As a scholar, I am very interested in AI, especially in focusing on what it can do, rather than what it can’t. In academia, many discussions about AI in education and research tend to be rather critical claiming that AI can never be “trustworthy” or used for complex tasks where the end result is important. I totally agree with that but in many cases where AI has rendered itself useless, it is often due to a) poor prompting, b) attempts to use it for tasks it was not developed for, and c) unrealistic expectations.

The text you are reading right now is not generated by AI, but like much else on this blog, it has been edited by ChatGPT 4.0. The reasons for this are quite obvious: I am not a native speaker of English and I would prefer not to sound like an idiot in text. For this purpose, ChatGPT actually works quite well, but I also need to be careful about what I prompt into it and what I expect from it. I usually make a prompt stating that I only want to focus on grammar and spelling, without losing my tone or style. After that I paste the text into ChatGPT and review its suggested changes. I often choose to use its grammar and spelling suggestions, but I frequently have to re-do its use of commas and similar things, since these do not feel natural to me. I thus use AI as a smarter version of Word, not to save time (though it does save me time) but rather to polish a product I have already written.

Another way I use AI is as a smart search engine. For example, I often ask AI about texts I once read but have forgotten how to find or I use it to locate material on the internet that relates to my research. A while ago, for instance, I needed a text on why certain religious artefacts are not supposed to be conveyed to people outside a belief system, which makes them ethically unsuitable to be placed in a digital collection. ChatGPT identified a number of texts with this reasoning, and I chose one which I then carefully read. Hence, I did not “buy” ChatGPT’s own interpretation, but used it to find sources suitable for the task that I could later double-check.

A third way I use AI is to identify publishing opportunities. I usually describe the research I am about to do, and it helps me find a suitable journal after some time. Once the journal has been identified, I also ask it how articles in that field are generally structured, what sort of reference system they use, and so on—thus helping me build the formalia and structure. I usually double-check this with a selection of actual articles.

To summarize: AI is useful, but not in the way a lot of people use it. You still have to be prepared to do the work yourself, but it can give you the tools to carry it out.

Comments

Post a Comment

Popular posts from this blog

Contemporary Sexual Politics: Efforts to Silence Sexuality in Politics

Image
Following up on last week’s post, we can ask ourselves where questions of sexuality are being challenged in politics. From a contemporary perspective, I would argue that we are currently witnessing global attempts—particularly from conservative actors—to banish issues of sexuality from public discourse. In Foucauldian theory, we can speak of several forms of the exercise of power. Foucault uses the concept of biopower to analyze social views on sex, capitalism, and power. One of the central aspects of biopower is that it often operates through the individual, who is expected to internalize social norms rather than be externally forced to accept them. Yet, in the case of sexuality, we have over the past decades witnessed rather forceful attempts to silence it, thereby making people unaware of important aspects of their own sexuality. In such cases, the effort seems to lie in banishing "unwanted" expressions from public discourse, thereby limiting the individual’s scope for a...
Read more

The old bucket and me - Hinke Bergegrens philosophical works

Image
(the title of this entry is a play with words, since the name "Hinke" carries resemblance to the word hink in swedish which means bucket).  For the past months I have been working on a project on Swedish politician Henrik "Hinke" Bergegren, who is often forgotten in contemporary history writing. Bergegren was born in an upperclass family in Stockholm in the year of 1861 and early in his life came into contact with rising workers movement. Within the movement he soon gained prominence and was a key person in forming the social democratic party, which would shape Sweden upuntil this day. In the scholarship on Bergegren a key concern have been the framing of him as a political terrorist, where he on the party's congress spoke openly about using violence as a mean to achieve the workers goals. In 1908 he was excluded permantly from the social-democratic party, which often have been described as a result of his agitation for violence. However, historian Åsa Linderbor...
Read more

Book review of sorts: A history of childhood by Colin Heywood

Image
One of the many perks with being a university teacher is that you get to read alot of various scientific literature. This semester I have been assigned to teach course about the history of childhood, which actually aligns rather well with my research interests on the history of families and migrant children. However, teaching also requires to read through the course literature and in this case a key-book was Colin Heywood's "a history of childhood".  As a standalone book it works rather well as an introduction to both the history of childhood and, in a wider sense, social constructionism. What I however have struggled with when incoorporating it to teaching is these three aspects: a) The structure which is divided thematically rather than chronologically  b) The books overestimation of the social; whilst also underestimating human evolution c) The approach where the past is presented as a one-way route to an enlighted, contemporary Europe or western world  Starting with m...
Read more