Atlassian hits the AI wall
The co-founders of Atlassian, both long time advocates it seems of complete tech freedom from regulation, are now confronted with AI seriously threatening their business. What are we to make of this when we are considering policy for AI and big tech? The answer depends on whether we are seeing genuine “creative destruction”, and so productivity gains, or an exercise mainly in economic rent extraction.
Atlassian’s CEO Mike Cannon-Brooks last week announced that around 10% of the firm’s workforce would lose their jobs. Clearly AI can allow Atlassian to create its products with less staff, so AI can be a benefit. But with the share price down around 80% since its peak the share market seems to be telling Atlassian something else: using AI, its clients can build their own like products and, indeed, may need less of them.
This logic has, and will in future, see job losses and share price declines in many other firms. Should we be concerned? On the contrary, we should applaud what seems on its face to be clear “creative destruction” as the old way of doing things is replaced by the new. Productivity growth is the outcome.
People will lose their jobs which is a real blow for them, but just as with the massive tariff cuts late last century different jobs will be created, and Australia’s economy will be better off. Indeed, Australia currently faces a labour shortage in key areas as we need more, say, geoscience and other science-based graduates and skilled blue collar workers. Those choosing their careers will progressively understand where the future jobs lie.
From an economic point of view the digital economy has brought both creative destruction and rent extraction. Digital cameras replacing kodak films is one of countless examples of the former. But digital advertising or ad tech is now effectively controlled by Google which extracts enormous wealth. Likewise, Google and Apple App stores. In so many parts of our economy big tech has inserted itself between providers and consumers and extracted profits to the cost of consumers, product providers and the economy. When a small number of large companies control areas of the digital ecosystem the potential economic gains are throttled.
Atlassian’s co-founders have been strong supporters of allowing big tech to do whatever it wants. For example, Mike Cannon-Brooks has described Australia’s News Media Bargaining Code (NMBC) as a “shakedown” of tech companies and a way to extract money from digital platforms like Google and Meta. Stunningly, no mention from him of the clear “shakedown” in Ad tech and with App stores. Worse, the NMBC simply sought to preserve journalism by seeking compensation for the use of their content. Google and Meta do not produce journalism; no creative destruction here as no better “mousetrap” is being invented.
Mr Cannon-Brooks has also argued that the NMBC misunderstands how the internet works with “free linking”. Try telling this to companies using Google ad tech, or banks who cannot compete with Apple Pay because they cannot gain access to the NFC antenna in iPhones. Nothing seems free when it comes to the internet.
His co-founder, Scott Farquhar, heads the Tech Council of Australia who have been leading the charge on big tech’s push to have a more generous definition of “fair use” in copyright law so that AI can continually use media and various creative content without paying for it. No creative destruction here; more like theft.
The 2024 recipients of the Nobel Prize in Economics, Daron Acemoglu and Simon Johnson, have argued that the benefits of new technologies are not automatic, but rather depend heavily on the structure of the economy. For example, if technology is implemented within "extractive" institutions—structures that benefit a few at the expense of many—benefits will be concentrated, increasing inequality, as they argued has happened over the past 40 years. Alternatively put, we get good outcomes with competition, but with monopoly or oligopoly we usually get wealth extraction.
Australia needs some basic rules of the AI road. For example, markets cannot remain free while conflicts of interest are allowed to fester such as when a company can both own a platform and compete upon it.
In his speech at the Australian Governance Summit, Commonwealth Bank of Australia Chair Paul O’Malley was reported as speaking about the “geo-economic” risks posed by the dominant US AI companies if they are not properly taxed and regulated. “Artificial intelligence has the potential to lift productivity across the economy, but only if Australia captures the value,” Mr O’Malley was also reported to have said.
Mr Cannon-Brooks and Mr Farquhar are not calling for any regulation to prevent damage to Atlassian, and we certainly do not need this. But contrary to their general approach to AI and big Tech we do need some basic “rules of the road’ to ensure AI boosts productivity rather than creates more opportunities for rent extraction.
This article was first published by the Australian Financial Review and is authored by Rod Sims, Enterprise Professor, Melbourne Institute of Applied Economic and Social Research, University of Melbourne. From 2011-2022, he was Chair of the Australian Competition and Consumer Commission.