I spent most of last week at SXSW Sydney, so of course I made a pitstop at the Darling Harbour branch of SMiZE & DREAM, the ice cream shop owned by SXSW Sydney’s own Tyra Banks.
While handing me a scoop of surprisingly decent vegan ice cream, the woman behind the counter clocked my lanyard and asked what I’d seen so far.
“Well,” I said, “there sure are a LOT of panels about AI this year.”
“Oh, I hate AI,” she said. “I want to be an actress, and now they have this AI actress taking roles that I could play. There are so many actors looking for work. It’s a solution to a problem that doesn’t exist.”
I was very much in agreement with all of this. But, perhaps worried that I was some sort of AI evangelist, and wanting to smooth over any offence, she softened her stance.
“I think there are probably some situations where it’s helpful, and some situations where it isn’t,” she said. “As long as we use it the right way and not the wrong way, I think it’s okay.”
“I mean, yeah,” I replied. “But I guess the tricky part is that we all have to agree on what the right ways and the wrong ways are.”
Back in the SXSW halls, the AI saturation was real. Panels on productivity, panels on ethics, panels on creativity, more than a few panels with titles and summaries that sounded like they were generated by AI.
Everybody felt the need to stake a position. Even ‘Bridesmaids’ director Paul Feig, in town to accept his SXSW Screen Festival Pioneer Award, felt compelled to weigh in during his keynote.
“I’m not worried about AI,” he said, “because all AI can do is recycle what’s been written already. But everybody here has gone through something that’s unique to you.”
Dr Sandra Peter and Dr Kai Riemer were somewhat less Pollyanna-ish, warning that at a time when 90 per cent of data online was created in the last two years, we are “drowning in slop”.
The University of Sydney Business School academics used their session to launch their 2026 Skills Horizon report. They made the case that we already have all the ingredients for a new era of work, including Gen AI, but nobody knows quite how they’re meant to fit together. “All of us,” they said, “are in R&D now.”
(That’s why Ryan Foutty, VP of Business at Perplexity VP, told attendees of his session that the people who are successful in this era “won’t necessarily have all the answers, but they’ll ask the best questions”.)
Drs Peter and Riemer acknowledged a projected rise in productivity from Gen AI use, but also showed research on what happens when we offload too much thinking. Brain activity drops. Expertise flattens. Workplaces fill with vibers; people performing competence, rather than developing it.
This is a good deal for some, they said. People in senior roles, with context and memory of the ‘slow’ way, get sharper. Juniors, without that grounding, get faster but dumber.
In another session, Dr Michelle Dickinson took that anxiety and said the quiet part out loud. “AI should only be given to old people,” she said, proposing a minimum age of 35 for a Gen AI licence.
“Young people don’t see the shit that AI writes that is just nonsense, because they haven’t been in the game long enough.”
But it doesn’t seem like opting out is really an option anymore. Sandy Carter, CEO of Unstoppable Domains and one of the more influential voices on the subject, made the case that AI adoption is inevitable, but thinking in terms of “fast” and “slow” isn’t quite it.
“Don’t think of it in terms of, ‘How do I use AI to do my current job faster?’” she said. “Instead, ask yourself, ‘What could my role become when AI handles routine execution?’”
If AI takes care of the doing, then the leverage moves to deciding. That’s why, in Drs Peter and Riemer’s view, the biggest advantage in the workplace of the future will be taste.
Knowing when to trust your own eye over the machine’s; knowing where the line sits between automation and imagination; knowing what to accept and what to reject.
From now on, we are all John West.
So after a week of panels, provocations and palpable feelings of existential dread, I arrived at one inescapable conclusion.
I think there are probably some situations where AI is helpful, and some situations where it isn’t. As long as we use it the right way and not the wrong way, I think it’s okay.
