People tell me stuff. Often they tell me stuff that is confidential.
I have been thinking about AI and privacy, and although this is unlikely to be an articulate and polished piece of prose, it is something I have been dwelling on for a while. So, as a priest in the Church of England, I think people expect to be able to tell me things without fearing I’ll gossip – and I don’t.
There was a TikTok circulating recently about a pro-Russian politician asking (in some European parliament) why google advertises child porn. The reply was, “it’s based on your search history”. Ouch. I think the term my kids use is pwned. Many (many, many, many) years ago I spotted that google was reading my emails. Whether anyone at google was reading my emails was moot, it was obvious that google was offering to auto-text reply to private emails based on their content. Way-back-when I asked myself if this was a confidentiality problem, and concluded that I didn’t know… however, protonmail was making a name for itself as WikiLeak’s confidential email provider of choice and I figured if it was good enough for WikiLeaks, it would be good enough for a parish priest. I opened a paid-for account.
More recently, with the rise of generative AI, I have become interested in copyright and the way AI gouges human-human interaction in order to fake it back to us. Meta’s AI trained on facebook doesn’t know how the printer works and writes angry letters to local newspapers. ChatGPT hallucinates. Grok tweets encouragement to fascists. All of this is classic human-human behaviour, and nothing to be worried about. However, the data comes from somewhere, and while I am deffo bothered about copyright, I’m also more bothered about confidentiality.
Most people are lackadaisical about confidentiality, or really angry about it. Or both. I used to work with statistics generated by global population health data – these datasets told me nothing about your Mum’s hip operation, but gave me insights to disparities of health care between different ethnic groups. However, if a national newspaper ran an article about ‘big pharma’ having access to ‘your health data’, there was usually some induced outrage. People are stupid. I say that knowing I’m people too, but this concern about ‘big data’ really is stupid: this big data isn’t a breach of confidentiality and it might help us all live healthier lives.
No: the confidentiality I’m bothered about is more personal. We’re not talking about the seal of the confessional here, rather the same confidentiality you might expect from any professional.
If a parishioner emails me about a personal matter, I understand they expect confidentiality. My non-judgemental openness to humanity’s brokenness is limitless. I hear much, and I hold it before God in prayer. Spoiler: I’m not telling tales to God… God already knew. However, I don’t think my parishioners expect adverts about it on facebook, or google search results for remedies to their personal problems. Yet this is the level of AI evesdropping that is currently going on. My smartphone asks if I would like a summary of PDFs I’ve been sent – the assumption being I’m too lazy to read. Google and Microsoft are offering me insights and summaries to my emails already – I even have a Diocese of Leeds email address, but I don’t know where the email server is or who has access…
So I’m sticking with protonmail (ymmv) and then, if you email me from an unsecure server, at least any indiscretion isn’t mine.
Or, perhaps send me a letter. I know the postie doesn’t read those.