Published on the 14/11/2024 | Written by Heather Wright
Time to get real on AI benefits…
“AI is an amazing technology shift… but let’s not pretend it’s going to double our productivity as a nation. Capability does not necessarily mean applicability.”
Brett Roberts is open that he has a bee in his bonnet about claims being made around the benefits of AI for the country’s productivity.
“A little bit of AI sprinkled over the top is not going to double our productivity,”
The tech sector veteran – he’s held senior roles at Microsoft, Datacom, Callaghan Innovation, the Icehouse and Air New Zealand (as IT manager in the early 1990s) – says while he’s never seen anything as ‘massive and profound’ as the explosion of large language models such as ChatGPT, he’s concerned about AI-washing.
An Accenture report for Microsoft earlier this year claimed generative AI was expected to more than double New Zealand’s productivity. The Australian version of the report, this time for Microsoft and the Tech Council of Australia, also made bold claims about the impact on the Australian economy and productivity.
Roberts is scathing of the reports, saying to talk about the potential for productivity to be doubled is ‘the most outrageously stupid, incorrect and factually nebulous comment’.
To be fair, the reports did also highlight that those gains were unlikely unless certain conditions were met – including bolstering woeful digital maturity in New Zealand’s case – and appear in some ways to be designed to be a call to arms for more focus on developing the policies, skills and trust needed to reap the full benefits of the technology.
“How much of this potential value is captured depends on the pace at which generative AI is adopted across all industries and how well workers are supported to transition to other tasks,” the Australian report notes.
And therein lies the crux of the matter in Roberts’ eyes, with the bulk of New Zealand and Australian companies being small businesses – among the least likely to gain huge wins from GenAI and propel A/NZ’s woeful productivity up.
“In the 40 years between 1978 and 2018, the GDP per capita gap between Australia and New Zealand grew from 14 percent to 41 percent – we are going backwards quite spectacularly as a country, productivity-wise,” Roberts says.
It was in the midst of that decline that the IBM PC was released.
“The greatest, most fundamental and profound technological change in our lifetime hasn’t arrested that decline. Maybe it slowed it, but our decline in productivity has been ongoing for a very long time now and is underpinned by a bunch of other far more major structural issues within our country.
“A little bit of AI sprinkled over the top is not going to double our productivity,” he told iStart.
“Things like [the productivity claims] get my goat for a couple of reasons. One is it is terrible hype and anyone who can run a spreadsheet would know it is not going to happen, but the other thing is I have this fear about credibility damage.
“I worry there are folks within government and other places who understand economies, economics and productivity, who look at comments like that being made by the ‘industry’, dismiss them outright and look askance at any future comments made.”
A Datacom paper has also talked up AI’s impact, saying 80 percent of Kiwi business leaders say AI is having a positive impact on business operations. That paper is focused squarely on large 100+ employee organisations.
Roberts doesn’t have an issue with that report, so long as everyone is clear that it’s only relevant for the ‘tiniest percentage’ of local companies.
“We have to be honest and practical about it.”
To be clear, Roberts is a believer in AI. He believes it will have a profound impact on some segments, including health, engineering, accounting and legal – all of whom have swathes of data and information built up over a long period of time, and where jobs are about mining that information or joining parts of it together.
“I do believe technology is a way out of us being a low productivity nation, but throwing vast amounts of AI at things isn’t the way to do it. There are certain pockets and places where it is going to be transformational, but the reality is the average number of people in a business in New Zealand is 4.3 – the same as it was 10 years ago. Ninety-six percent of businesses in New Zealand have 20 or less people, and those folks probably still struggle putting a pivot table together in Excel.
“I remain to be convinced that they are somehow magically going to pick up AI and do magical things with it and transform their productivity.”
Which brings us to another bugbear for Roberts. He’s calling on the tech sector to take a good long look at itself in the mirror and acknowledge that it doesn’t do the best job at ensuring customers get the most value from technology they already have – and address that issue.
“We are always pretty good at the next new thing – that’s how sales quotas work – but I don’t know if every company on the planet has extricated as much as they possibly could from the stuff they already have,” he says.
“As an industry we should not talk about things like shifting the country’s productivity like that without some really hard data to back it up and some plans to do so.”
So what needs to be done so smaller businesses can truly benefit from genAI?
Roberts admits he doesn’t know the answer, though improving digital literacy – as pushed in the Accenture report – is a critical issue the country has not yet managed to successfully address.
While most will wind up with GenAI tools – trojan horsed into their business through the likes of Microsoft’s Copilot offerings, Roberts is unsure if they’ll actually be able to use the to really benefit, without assistance.
He’s keen to see a long-term – we’re talking 10+ years here – bipartisan effort to lift digital skills of business owners and workers across the country, and notes US programs providing loans to help small businesses build out technology.
But he’s equally keen to see the government invest the time and energy into working out where AI can truly make a difference for the country’s productivity and economy.
“We have to work out those places to invest, put real tangible plans together and then go and do them.”
And that requires cutting through the hype that AI is going to change everything for everybody.”