American technology journalist and author Jacob Ward says AI is accelerating faster than our understanding of human behaviour – and the business incentives behind it risk “flattening” human choice, creativity and agency.
“It’s this incredibly turbocharged vehicle – and at the wheel are a bunch of people, myself included, who have no idea how to drive,” he told Commercial Disco.
Mr Ward describes the AI moment Australia is stepping into, arguing that a collision between behavioural science and commercial AI incentives has created a system designed to exploit the most primitive parts of the human brain, not elevate them.
According to him, the last decade of behavioural research shows how easily predictable (and manipulable) human decisions are, while the tech industry is simultaneously deploying tools capable of shaping behaviour at unprecedented scale.
The danger, he said, is not just the technology but the incentives behind it.
“Companies spending more money than any in history now have to make that money back,” he said. “And they’re dealing in who we think we want to be versus who we really are.”
One of the most striking examples Mr Ward describes comes from inside OpenAI, whose researchers previously wrestled with an internal ethics dilemma they called “the heroin problem.”
If an AI assistant is supposed to optimise your life – but you’re addicted – should it help you score with greater efficiency, or intervene?
“They didn’t know the answer,” Mr Ward said. “Do you let people choose their own path, or do you take a paternalistic view of what a good life looks like?”
The tension between prediction and autonomy now sits at the core of every major AI design decision, with an emerging business model built on exploiting “revealed preferences”, according to Mr Ward.
Revealed preferences are the unconscious impulses that social platforms have learned to capitalise on during the last decade of engagement-driven design.
“The like button was a primitive version of it; AI is the industrial-strength upgrade,” he said.
For all the alarm, Ward is not an absolutist. He is animated by the potential for scientists, researchers and public-interest technologists to use AI for genuine social good.
“Give it to scientists,” he said. “I want archaeologists, social scientists – the people doing the hard stuff – to go hard at this.”
But he is clear-eyed about where the money is not. “Nobody’s getting rich decoding whales,” he said.
Instead, the commercial frontier is already accelerating toward emotionally manipulative experiences: “AI companions,” as OpenAI now explicitly frames them and efficiency tools that risk hollowing out nuance, originality and thought.
Citing new research, Mr Ward said students using Google to gather health information produced more varied and nuanced writing than those who used AI to generate the same content.
“The AI responses flatten everything into a kind of anodyne, generic advice,” he said.
Mr Ward believes the regulatory frontier will shift toward something he calls cognitive liberty – the right not to have your decision-making manipulated by predictive behavioural systems.
Despite his warnings, Mr Ward ended on a surprisingly optimistic note, placing his hope not in regulators, but in cultural backlash.
“American teenagers have already coined a term for people who lean too hard on AI: clankers.”
“Kids are looking at us using AI the way we looked at our grandparents smoking at the dinner table. That’s what’s going to save us,” he said.
Do you know more? Contact James Riley via Email.