Alan Finkel is the former Chief Scientist of Australia and founder of the global certification organisation Proudly Human which aims to protect the integrity of human creative endeavours. He spoke with Steve Crossan who was a founding member of the pioneering DeepMind project, and is now actively engaged at the interÂsecÂtion of AI and biotech.
For both, the potential rate of development of AI is something that is going to have huge implications for society.
“I was always very, very comfortable with what is actually a fast rate of development previous technologies,” said Finkel. “There’s a law called Moore’s law. Gordon Moore, the founder of Intel, observed, that the capability of chips they were building was doubling every two years. The equivalent observation with AI is that it doubles every seven months, which means it’s a tenfold increase every two years. It’s cumulative, so if this keeps up in 2036, AI will be a thousand times more powerful than it is today.”
Crossan is also very interested  in the emergence of AIs that understand code as well or better than they understand human language.
“Around December 2025, we started getting systems which could reliably take on a task that a good human programmer would have taken a few hours to do. If it can’t do something, it can write a program to do the thing that it can’t do, which is an important step change.
There’s been some very interesting results in mathematics and physics recently. If you look at protein folding, until 2020 we had about 200,000 protein structures available to science, researched using x ray crystallography and very complicated processes- representing about 50 years of work from scientists all around the world. From AI working in the last 4-5 years, we now have well over a billion structures available to science. The impacts across all science domains is going to be huge.”
The talk moved on to the issue of AI and creative outputs. Finkel talked about human  provenance being important for society. Â
“I founded Proudly Human on the premise that I believe that there’s something special about creativity that we need to honor and preserve,” said Finkel. “It’s part of our social interaction with each other. We started with books and the idea was that an author, on the author’s own volition, will come to us, prove that they’re human, and they will be given a unique ID. There are a lot of things that are acceptable use of AI, like doing research, but it’s not acceptable for AI to write sentences, paragraphs and chapters. We do actually use AI powered tools to look for evidence of AI in the manuscript itself.”
“I can’t see the economic benefit in having AI write a book or produce a film. It’s just undermining human creativity and the essence of being a human.”
Crossan said it is important for authors to be able to stand by the thing that has their name on it.Â
“You have to deeply understand it, and be able to defend every word,” said Crossan. “If you can’t do that, you’re probably ethically in very reasonable trouble. And your audience has a right to understand the relationship that you have with the work.”
“We already have lots of examples where a CEO gives a speech they probably didn’t write, and a celebrity writes a biography they probably didn’t write. But we an expectation that the CEO or the celebrity would stand behind the words and would take responsibility if the words were found to be, erroneous in some way. So I think the same principle applies here.”
“I think the human being working as a speechwriter is acceptable, but I don’t think it’s acceptable for the AI to do that,” said Finkel, “because the ease with which the AI starts doing that is a credible threat and its just spread like crazy.”
“I would agree that it introduces different and new ethical questions,” said Crossan, “largely because of the scale at which it can happen and the scale at which it’s available. I don’t necessarily agree that a blanket ban is the right thing.”
I do strongly think that, other things being equal, having some way, in which we can as consumers or as readers or as audiences, understand more about the provenance of the thing we’re interacting with- that’s unequivocally a good thing,” said Crossan.
It’s an interesting tension between creativity, authorship and gatekeeping. If Finkel thinks it’s OK to use a human speechwriter, what happens to the majority of people that don’t have one. There is an element to AI use which is egalitarian and breaks through barriers of social and educational capital. Similarly with creative processes, modern and historical figures in the art world are littered with examples where the famous artist that collected the cash didn’t actually make all of the work. But maybe they made the original work which was then copied and replicated but assistants, students, and machines. It’s an irresistible conundrum which we will all be tasked with facing sooner or later.Â
