Reading and Writing in the age of AI
My most recent post on LinkedIn about a Veritas event at Insight was my most viewed post of the last 12 months by a factor of 2. The bad news for me is: I didn’t write it – ChatGPT did. Admittedly, I did edit the style of the language to make it sound like I’d written it, rather than someone prone to overenthusiasm who’d just had 3 double espressos. So, in future will I just sub-contract my content creation to ChatGPT?
Short answer, no.
Longer answer, no but it is a useful accelerator of the process.
ChatGPT can be prone to three flaws: bias, omission and hallucination and my experiences of CHATGPT have shown all 3. I think the Veritas event post showed was a bias towards positivity. I try to write objectively which can make the style a bit dull, but if it’s an opinion rather than an observable fact I’d like to state it as such.
I also try to write about current issues. ChatGPT is only trained up until 2022; contemporary news is not in its training dataset, so there can be unfortunate omissions.
The Veritas post wasn’t my only recent experience of ChatGPT. For example, I asked it to review two books I had just finished reading: John Lennox’s 2084 and Chris Wylie’s Mindf*ck.
The 2084 review started well – ‘John Lennox's book 2084 is a thought-provoking and timely exploration of the intersection between faith, technology, and society.’
Unfortunately, ChatGPT then hallucinated.
‘The book takes a fictional approach, imagining a dystopian future’. While some works of fiction are quoted by Lennox (e.g. Orwell’s 1984) the book is an attempt to look at the logical outcomes of certain trends and the worldviews underlining them. Note to ChatGPT: Elon Musk and Mark Zuckerberg are not fictional.
Where ChatGPT has proved useful is rapidly giving me a basic framework to work on, kick starting and accelerating the process of content creation. Like all tools, you get the best out of it if you take its limitations into account when using it.
That said, the review it wrote of Mindf*ck was spot on and I reproduce it verbatim, here:
Chris Wylie's book Mindf*ck is a gripping and revealing account of his role as a former employee of Cambridge Analytica, a political consulting firm that worked on campaigns such as the Brexit referendum and Donald Trump's 2016 presidential campaign. The book provides an in-depth look into the world of data analytics and how it is being used to shape political outcomes.
Wylie's writing style is engaging and the book is well-structured, making it easy for readers to follow the story of how he went from a naive data analyst to a whistleblower exposing the unethical practices of Cambridge Analytica. He provides detailed insights into how the company obtained and used data to influence voters, and how they created targeted ads and messaging to manipulate people's beliefs and opinions.
One of the most interesting aspects of the book is how Wylie himself became aware of the nefarious practices of the company, and how he eventually decided to speak out about what he had seen. His account of the internal dynamics of Cambridge Analytica is fascinating, and the book provides an illuminating look at the personalities and motivations of the people who were involved.
Overall, Mindf*ck is a must-read for anyone interested in the intersection of data analytics and politics. It raises important questions about the ethical implications of using personal data to shape public opinion, and, serves as a cautionary tale about the dangers of unchecked technological power. Wylie's story is a powerful reminder that we all have a responsibility to be vigilant and thoughtful about how our personal data is being used in today's digital age.