Modelled after machines: new ways of being human in the era of AI

Artificial intelligence’s accelerating ability to mimic traits we previously regarded as uniquely human has emerged into the popular consciousness in recent months, with a proliferation of pop culture applications of generative AI like ChatGPT and Midjourney.

The mimicry of the painting styles of great artists to absurd ends. The simulated superstars, the feuding bands reformed, the dead performers shocked back into life to cover other people’s hits. The inevitable pornography — both AI-generated images of real women created without their consent, and the creation of illusory women selling naked pictures. The potential for malign political uses has become apparent via images of an explosion at the Pentagon that never happened, and falsified footage of Ukrainian President Volodymyr Zelenskyy telling his people to surrender to invading Russian forces.  

As yet, these are easy enough to pick as frauds. The University of Wollongong put on an AI-penned play, and it was very bad.

“The language is quite generic and cliched, the characters are superficial and underdeveloped and all of the drama is sort of told to you as the audience,” University of Wollongong theatre fellow and the show’s director Matthew Prest told the ABC.

However, this is just the early stages. Venture capital firm Sequoia cheerily predicts that by the end of the decade, generative AI will write better than professional, and produce images “better” than “professional artists, designers and photographers”.

What will it do to our sense of reality and the way the world is organised to live alongside this increasingly ubiquitous, uncanny production of alternate reality that is hard to detect and subject to harmful commercial and political motivations?

Work and life

The first element is the most obvious, and the one for which we have the most historical precedent — what it does to societies when entire professions, and the communities built around them, are obliterated? The World Economic Forum predicts changes wrought by AI will eradicate 83 million jobs, and create just 69 million in the next five years. Around 26 million jobs in administrative positions will be cut due to AI. It’s an upheaval that has the potential to match that of the industrial revolution, which broke up families and wiped out communities as it sucked rural populations towards factory work and the cities.

“So there’s a huge shortfall of jobs created, but even if someone has their job replaced, it has implications for people who view their work as an important part of who they are,” Emeritus Professor of business and labour history at the University of Sydney Greg Patmore said, citing the example of a study he did on Nova Scotia steelworkers after their plant closed down in 2001.

“They were all promised new jobs and the jobs they were sent to were in a call centre. So it went from work that they had expertise in and felt pride in, work where they could see the outcomes of what they were doing, to a job where they felt none of that connection,” he told Crikey.

Call centres, incidentally, would count among the “administrative” roles that make up such a large part of the service economy in countries like Australia and are in danger of being wiped out by AI.

AI and reinforced reality

AI got a lot of attention last month when several of the architects of AI technology put out an open letter encouraging policymakers to view it through the same lens they view existential threats like climate change and nuclear war.

This points to what Dr Jenny L Davis, associate professor in the School of Sociology at the Australian National University, believes is the real lacuna in how we are currently discussing AI and related issues.

“The slate of what I call ‘AI doomers’ misdiagnoses the problem in some ways — absolutely we do need to slow down, we do need to fund more research,” she told Crikey. “But the main thing with large language models like ChatGTP is that they run on data, data from people, from us. So they will necessarily reflect societal bias and structural issues. If anything, they’re amplifying those issues by packaging our collective bias back to us as objective data.”

Reasserted humanity

If history is anything to go by, “there will be new forms of resistance” in the teeth of this, Patmore said.

“If you look at the industrial revolution, over that period trade unions and consumer collectives were greatly developed as a response to what was happening,” he said. “So currently, we’re seeing the emergence of things like data co-ops, people treating their personal data the same way groups did in the industrial revolution, as something they could collectivise and bargain with.”

Davis said large language models and large image models are, at their core, form without meaning or stochastic parrots.

“They are necessarily conservative, not necessarily politically, but simply based on where they get their information — that which already exists, and is subject to a lag of a few years,” she said.

Thus any crisis stemming from AI at this point cannot be new — by definition, the call is always coming from inside the house.

“AI is just repackaging our slightly older selves back to us,” Davis said. “We should be aware of that when we talk about AI as generative, and creative and novel. By definition, it can’t be.”

AI’s advances mirror the advances in film technology and recorded music in the early parts of the 20th century, said Robert Sparrow, professor of philosophy at the Monash Data Futures Institute.

“Recorded music essentially allowed people to approach perfection,” Sparrow told Crikey.

“The performance you hear isn’t a real performance, it’s a stitching together of the best performances, the vocals have been mixed — so it has a level of polish and perfection that no band could recreate. In opposition to that, particularly in forms like jazz, we see a fetishising of live performances, this notion of authenticity, imperfections, being in the moment,” he said.

“I suspect that we’ll end up both trying to become more the polished image of humanity presented by AI — we will model ourselves after our machines — and fetishising what we hold to be still distinctly human, including our flaws and imperfections.”

Davis agrees. “In many ways, AI offers a reassertion of the distinctiveness of humans, it articulates a sense of what AI can’t do that we can”.

Sparrow has a further technique, one familiar to the many people tired of the warping effect of social media: go outside for a bit.

“This will only be true online. AI has only advanced so much because we spend so much of our time online, and that part of our lives was already pretty shit,” he said. “In our daily lives we will still be able to get affirmation from other humans.”

Are you concerned about the rise of AI? Is your work at stake? Let us know your thoughts by writing to letters@crikey.com.au. Please include your full name to be considered for publication. We reserve the right to edit for length and clarity.