AI has already found some diabolical uses, from unconsented AI-generated porn to political misinformation-spouting bots. However, these are limited to individual rouge actors or openly crooked states. But I recently stumbled across an AI which, on the surface, seems legitimate, wholehearted and progressive. However, if you dig a little deeper, it’s actually profoundly horrific, and what’s worse, I don’t think the people using it realise that. You see, the Bletchley Park Museum recently announced they are building an AI-powered life-size model of Alan Turing, who can talk to visitors. I told you it seems innocent…
You’ve almost certainly heard the name, but let’s recap who Turing was.
Alan Turing was the father of modern computers. During WW2, as a part of the UK’s secret decoding team at Bletchley Park, he broke the German radio encryption machine known as Enigma by building and using a rudimentary computer from analogue switches and rotary drums. Since then, every computer has been routed in Turing’s original design for Bombe.
Turing had a promising career ahead of him after WW2, being a pioneering mathematician and a spearhead of the computing revolution and all. In fact, soon after WW2, he created the Turing test, a rudimentary test to establish if a machine has human-like intelligence. This is why some people claim he is the father of AI, despite the technology being invented decades after his death.
Sadly, it is what caused Turing’s untimely death and his unfortunate position in history which make this AI museum piece so damaging.
Alan Turing’s brilliant career was cut short in 1952 when he was charged with being a homosexual. He was given the option of prison time or chemical castration. He chose the latter, though it had some horrific side effects on him. What’s more, his security clearance was revoked, and as almost all of his work until this point had been top secret, this effectively stopped him from working. Turing was autistic, and this inability to work, combined with the brutalisation of his sexuality and body, caused a rapid onset of depression. In 1954, he died of cyanide poisoning, which an inquest determined was a suicide.
Turing’s work breaking the Enigma code potentially saved hundreds of thousands, if not millions of lives, by ending the war sooner. But even when Bletchley Park’s work was declassified in the 90s, Turing’s sexuality and abuse by the hand of the state were either overlooked or misrepresented.
As such, his place in history isn’t actually represented correctly. The record is skewed through this homophobic cultural lens, and while we do have diary excerpts from Turing, it is difficult to get an accurate depiction of his true experience.
The same is true for many LGBTQ+ figures in history. Indeed, you might not even know some historical figures were LGBTQ+. Their sexuality, a core part of who they were, and how they were persecuted, or even accepted for it, has been rewritten, swept under the rug, or twisted beyond belief to match the questionable morals our society had in the near past. For example, Julius Caesar had recorded sexual encounters with men, Alexander the Great likely had male partners, and there is even evidence Leonardo da Vinci was gay. Frida Kahlo, Billie Holiday and Josephine Baker all had a recorded history of same-sex relationships too, which have been twisted, mired or ignored in the historical record.
As such, this upcoming Alan Turing AI can only ever be an offensive derivative of our terrible ability to correctly record history.
You see, AI chatbots like this can only replicate the data they have been trained on and take a huge amount of training data to get even remotely accurate. There isn’t enough surviving contemporary material of someone like Turing talking about his experience to enable such an AI, and as such, it must be trained on secondary sources that have been distorted by this homophobic cultural lens. This means this museum piece will more than likely reflect the culture that suppressed Turing, not the man himself.
This is particularly relevant as the general public is now very aware of Turing’s sexuality, and as such, the TTP problem is very pressing. TTP is a video game development term, short for “Time To Penis”. In short, it means that if you give players creative freedom, there is a certain point at which it is guaranteed one of them will draw or enact a penis. How will this AI Turing respond when the museum goers with this creative freedom and ask it about sexuality and male genitalia? Programmers could block such prompts, but as shown with ChatGPT and other generative AIs, these blocks are easily bypassed. They could ensure it evades these questions or refuses to answer, but in doing so, it would be a huge disgrace to someone who was so horrifically persecuted for their sexuality and be a major ethical breach. Likewise, letting the AI talk, which will have very little training data on Alan’s true feelings about his sexuality, could easily lead to it going off the hook and delivering random, illogical and potentially harmful answers, which is equally ethically bad.
What’s more, Turing had his voice taken away from him. He was marginalised and abused to the point of suicide, and as such, couldn’t partake in the telling of his own history. To talk on his behalf, even if you do so tastefully, represents a major breach of ethics and twisting of history.
Now, I should state I do not believe Bletchley Park is proactively doing something damaging here, and I wish them no ill will. Turing is their figurehead, after all. He has also been inaccurately linked to AI for years, and the technology has a vast cultural buzz around it. So, such a museum piece makes excellent sense from a marketing point of view. I just don’t think they have taken the time to consider the limitations of AI and how such a piece would continue our society’s propensity to steel LGBTQ+ voices, undermine their life experience, and twist their true nature to the point of cruelty. But even if Turing didn’t have this position in LGBTQ+ history, resurrecting the dead with this cheap and often inaccurate parlour trick is deeply unethical. We cannot speak for the dead; trying to will only be deleterious. Instead, we should let the dead rest in peace, take the time to understand that our record history is inherently unreliable, and understand that AI has some major flaws we must mitigate.
Thanks for reading! Content like this doesn’t happen without your support. So, if you want to see more like this, don’t forget to Subscribe and help get the word out by hitting the share button below.
Thanks Will, for this sensitive and, I’m sure, correct anticipation of the problem of this misguided project. Alan Turing was undoubtedly a very great man regardless of the mores of his time, and their tragic outcome for him. Given the many issues of AI at this early point, it would not serve such a complex person and situation well. A sensitive biography like Andrew Hodges’ would serve an interested enquirer much better.