I don’t garden, but if I did, I can only imagine it would be like emptying my inbox. You might think weeding out emails gets easier over time, but it remains a tedious and remarkably painful task. Unlike actual weeds, which at least sprout a delightful yellow flower, I am perpetually disappointed by the notices informing me that I have not received a free cruise and my car warranty has expired (full disclaimer: I do not know how to drive nor possess a licence.) I am not a fan of this kind of weed, which is to say, there is a reason I do not garden.
Even disregarding spam, answering emails remains an unglamorous offshoot of young adult life. From brainstorming the perfect diplomatically passive-aggressive follow-up, to deliberating the appropriate number of exclamation marks to convey your moderate but not concerningly excessive enthusiasm for a rote assignment, we are all inconvenienced by this time draining requisite for students and early career professionals. Still, as many in this modern age have forgone the watering can in favour of sprinklers, we can also banish this dull activity and fulfil our inferred collective goal to become the sentient lumps from Pixar classic Wall-E. We, as a society, have progressed past the need to twiddle our thumbs thanks to the digital hands of artificial intelligence (AI) coming to plant an inbox near you.
In all seriousness, the debate over artificial intelligence doesn’t have to be a dull affair. While posed to augment our productivity for better or worse, it is unmistakably here to stay. Thus, rather than muttering vague prophecies of doom, we would be better suited to accept our strange reality with our senses of humour intact. In their new show ‘Who Wants to Write an Email?’ showing from 9th-12th September at the Lir Academy, Laura Allcorn and Jennifer Edmond aim to familiarise us with our brave new world through the unusual lens of a Jeopardy!-style game show. I spoke with Professor Edmond earlier this month, an Associate Professor in the Digital Humanities who graciously agreed to an interview.
“Rather than muttering vague prophecies of doom, we would be better suited to accept our strange reality with our senses of humour intact.”
Our conversation explored the linkage between art and the latest AI technology. Although many consider email writing more of a chore than artistic hobby, Professor Edmond concedes that there is an artistic element to writing emails that may be lost when AI takes over. “Every time we adopt a new technology we gain a lot, but we lose a lot as well. And we’re usually very cognizant of the things we’ve gained, but we’re often less cognizant of many things we lose.”
Edmond notes that the new technology itself shapes the discourse of connection. You have likely already noticed this tendency, shaking your head at the meticulously capitalised and punctuated paragraphs sent by older relatives unfamiliar with our chronically online generation’s texting habits. The same applies to a bygone era when prehistoric users of the artefacts known as ‘flip phones’ had to “push the button three times in order to get certain letters” and navigate a less capable processing system. This anecdote of the olden days sounds even more painful than dropping a gardening shovel on one’s foot.
“Edmond notes that the new technology itself shapes the discourse of connection.”
While the physical planet has been fully explored, new technology necessitates the birth of a new language for navigating this unfamiliar landscape. According to Edmond: “What happens is usually culture catches up. And it’s really interesting to talk to people about SMS messaging now because there is a whole very clear [language where] if that’s capitalised and there’s a full stop there, then that’s really aggressive. But if you don’t capitalise it, it’s fine.” Her reasoning unpacks a phenomenon we all too often take for granted. Indeed, it is only with the advent of text messaging that the period has become an effective declaration of war. Proper grammar now feels suspicious, the exclusive province of shady characters like your grandma. When my friends use fully capitalised sentences I instinctively suspect they are plotting my downfall. I am relieved to hear from an expert that my suspicions have linguistic origins.
Along with the territory, we discuss the dangers inherent to the dreaded email. Similar to texting, email-writing entails losing “so many of the signals we’re used to having around communication”. Edmond explains: “You lose tone of voice, you lose facial expressions, some of which were maintained by certain technologies” such as the telephone which allowed users to retain their tone. The virtual impoverishment of signifiers can lead to misunderstandings. “I know we’ve all had the experience of writing an email where we just kind of send it and then the response we get back tells us that it just didn’t sound to the recipient the way it sounded to us because of that impoverishment.” This road has certainly been travelled down many times and due to the language barrier, it is difficult to avoid when you are “just dealing with the words on the screen”.
Artificially intelligent email writing should be no surprise, considering spell-checking tools have been stealthily infiltrating our writing for years. Spell-checkers, and the suggestion tools of platforms like Microsoft Word, have linguistic quirks of their own. Their tendency to favour certain words over others influences our writing style. For Edmond, the implicit bias feels like an encroachment on her self-expression. Although more academic writing must conform to strict standards, personal communication can be an outlet for creativity and when entire emails are written by an AI agent, this innovation is stripped from us. It is a more extreme example of how AI affects the discourse as it replaces us entirely. Since machines can replicate our language, the person on the receiving end may not realise they are talking to a robot.
“Personal communication can be an outlet for creativity and when entire emails are written by an AI agent, this innovation is stripped from us.”
During the show’s testing process, Allcorn and Edmond mused about the proper balance of communicating with human flaws and regulating AI communication to mitigate the cultural risk. One crucial take away from the co-creation events with potential audience members was an understanding of certain events where AI email generation would never be acceptable. Edmond says: “a sympathy note should never ever be written by an AI. You had to write that yourself. And even if it was difficult, even if it was hard, it’s a very human hard, and that’s something that unless it comes from the heart, it’s absolutely meaningless.” AI can also be quite awkward in such situations, as Edmond notes that when a system called Hugging Chat was called on to respond to “a very personal interaction” it failed miserably to the point of using “this sort of officious business language”. The takeaway is: don’t ask artificial intelligence to write your condolences because like Patrick Bateman excusing himself to return some videotapes, it will make you sound positively sociopathic.
Edmond found it interesting that although the potential contestants rooted for a human interaction, the machine was incapable of generating the same outpouring of emotion. She muses about a possible dystopia we must work to circumvent: “ I think that even though sometimes this sort of communication is hard for us, that hard is good for us, how we learn, that’s how we grow, that’s how we connect, that’s how we create social capital. And if we were to get to… a world where we actually don’t read our own emails anymore because my agent writes to your agent and your agent writes back to my agent, we’d become detached from our own sociality, alienated from it. And that I think is really, that for me is the sort of, the scary future. But of course, that’s also a place where culture is going to invent some other way of communicating.” In speaking of agents, Edmond evokes a future that recalls my own memories of interacting with large corporations, where to get to a top executive one has to be vetted by three communications teams just to make it to the inbox. The vetting effect ensures a lowly employee will never be able to make their voice heard. I suspect the inevitable reality of this AI agent-based communication is to define a new world where one’s humanity is irrelevant; where one operates at a complete impasse from imperfect – and authentic – phrases capable of eliciting discomfort; where spell-check is applied on the level of emotions.
“I suspect the inevitable reality of this AI agent-based communication is to define a new world where one’s humanity is irrelevant.”
That said, my speculation may be likened to a toothless old woman on her lawn prophesying doom for kids these days, and in the strangeness of new technology, humour is also evident. I asked Edmond whether she thinks comedy can help people wrap their heads around emerging technology, and she replied in the affirmative. Artificial intelligence is frightening in its unfamiliarity in contrast to ‘Who Wants to Write an Email?’ which, as a gameshow, is both familiar and hilarious for its absurdity. Edmond praises Allcorn, the founder of the Institute for Comedic Inquiry and partner in developing the game show, for teaching her about “helping people to break through that dichotomy… either it’s like, we can’t criticise it [because it is ] so important because [of] the future or it’s horrible and we shouldn’t get anywhere near it. And of course, the reality is in between the two, we are going to use it. We do use it every day whether we know it or not. So the more we can understand it and the more we can feel empowered in the face of it, the better.”
Comedy, the art of illustrating the absurdities in life, is also a mechanism for ingraining cultural ideas in the audience. Game shows are also a traditional hallmark of societal norms and trends, showing the audience aspirational models for success in the forms of prizes, life goals and those pop culture references that contestants take pride in knowing. They are also incredibly popular, particularly among older audience members. Therefore, the game show format is a natural fit to encourage users who would likely refrain from using AI to adapt to it in a productive manner. The transcript of a gameshow is formulaic but, for viewers, the experience is not. The same applies to AI, where Finland came up with a dry, instructive course that “was very serious, it was very earnest, it was very Finnish, but that doesn’t really work culturally everywhere and it doesn’t work for everyone.”
Explaining how the show operates, Edmond elaborates that the combination of familiar trivia knowledge and randomness where you have “these four choices and you can guess” ensures that the show remains lighthearted. The central messaging is “we’re all the humans in this together and we’re all trying to figure this out together. But at the same time, we want people to walk away saying, ‘Okay, I kind of know my comfort level now with this, and I feel like I can confidently say that I’m okay with it there, but I’m never going to consider writing a consultants letter, for example, using AI.’ ”
“The central messaging is ‘we’re all the humans in this together and we’re all trying to figure this out together.'”
What about AI’s capabilities to monopolise art? Researching the potential for artificial intelligence through the Center with Digital Humanities, Edmond cites the words of Marcus du Sautoy, a journalist, researcher and mathematician who wrote The Creativity Code. She says: “AI can create art, but it can’t have the intentionality behind it. And if you think about it, art is a product. We can view art as a product. You can buy a poster to hang on your wall, you can pay Spotify to be able to listen to music. So there is that product aspect of it. But really what’s more interesting is the process aspect of it is the reason why we engage with art, why it makes us feel something, why we want to create and that’s something a computer will never have.”
Furthermore, Edmond notes that artificial intelligence is artificial – that is, inferior to our natural process – because it is dependent on human experience. In turn, human knowledge and our “conditions of culture” also constantly change. Artificial intelligence will always play a reactive role in response to cultural evolution and will thus always “lag behind” and never be able to reach a new state by itself. “I think that AI can imitate a lot of what we do as humans, and it can imitate it very well and can imitate it in ways that bring together new configurations of words and new configurations of ideas. But it takes a human to say from that, oh, this gives me reason to live. This gets me past my existential crisis is something that is meaningful and that needs to be passed on to the next generation.” There cannot be existential rebellion to the artificial intelligent agent which fundamentally accepts the conditions and terms of the society it has emerged within, that are coded into its very being.
Instead of looking to artificial intelligence as a replacement for creative occupations, which Edmond defines as a capitalistic impulse (“if we could just get the humans out of the way, we wouldn’t have to pay them,”) she thinks the goal should be attaining a “sweet spot where… you have that sort of augmented intelligence where we’re using, we’re not just saying, oh, the machines can do it all”. For instance, artificial intelligence can augment human traits such as the “sensitivity around empathy” that many doctors are criticised for and use generative tools to critique and improve their communicative skills. Becoming better at communicating is the imperative rather than having RoboCop prescribe your antidepressants and perform open heart surgery.
“Becoming better at communicating is the imperative rather than having RoboCop prescribe your antidepressants and perform open heart surgery.”
Ultimately, artificial intelligence will never replace us because socialisation is as essential as food or water to our human condition. In our modern world, we are prone to forgetting this fact but Edmond urges its importance, saying we need to consider how a replacement would harm our social interactions. “At the end of the day, we’re the mortals here. We’re the ones who need to be socialised so we can survive.” The cultural risks also have their life at stake; considering the ongoing Screen Actors Guild strikes. Without the human interaction on the page you lose the authenticity of the “original kind of voice” lacking from “that voice generated by a machine… AI doesn’t feel.” An ethical design policy must be implemented to “create an ecosystem that both allows creativity through these kinds of tools, but also protects the creators” through surveillance in fields like the visual arts to ensure training data doesn’t allow for the replication of a “very unique style”. Simultaneously, freedom in the artificial intelligence field is important because when training data is lost, so is an integral part of art – “that aspect of the collective chatter of humanity” – constantly reproducing and deriving its power from previous works. Sending an email to a friend, capturing the chatter of my internal thought process, suddenly seems more urgent than ever. Even if, like my propensity for gardening, talent does not come naturally.
Performances for ‘Who Wants to Write an Email?’ run at the Lir Academy – Studio 2 as part of the Dublin Fringe Festival from 10-12 September 20:45. Tickets can be purchased online at fringefest.com for €11-€13.
