Whatever the harm, it will be worth it – this was the basic message techno-futurist Eric Schmidt offered at a recent US senate hearing in April when he was asked to explain why he’d said that humanity must burn as much oil, gas and coal as it took to create the AI that would solve climate change.
“Knowing that climate change is real and knowing that it's a problem is that the intelligence revolution, the ability to do planning and discovery will allow us as Americans to develop new materials, new energy sources and so forth because of the AI data centres,” he said. “So our core argument is invest in the way we can now because the future will be so much cleaner and so much more efficient as a result of these algorithms.”
The former Google CEO was saying something maniacal, wrapped in tech exec logic. In his view, the catastrophic consequences that may occur from continuing to burn oil, gas and coal are justified if it leads to the creation of an AI system that will solve climate change, as is any damage done in the meantime. Schmidt is not alone in thinking this way. This technological accelerationism has increasingly been adopted by influential figures. Billionaires Bill Gates and Jeff Bezos have sunk hundreds of millions of dollars into developing AI as a “climate solution” and, a similarly credulous faith in technology to solve climate change has been trickling down through the discourse. This comes despite higher carbon emissions and water use associated with widespread adoption in the present. Meanwhile, the combination of massive energy demand from additional data centres and the use of AI by fossil fuel producers to extract more oil has led some researchers to describe the climate impact of AI as potentially “devastating” when “applied at scale”.
There is nothing new about this proposition. Before AI was being used by billionaires to justify delaying phasing out oil, gas and coal, company executives and political leaders were using economic modelling to argue any phaseout would be a disaster. Economic models, as sufficiently complex spreadsheets, can be thought of as an evolutionary ancestor to machine-learning algorithms and large language models. Bureaucrats and political leaders have long looked to modeling to ‘help make better decisions’, with climate change proving to be a convenient laboratory for these innovations. In hindsight, however, the use of these spreadsheets appear to have served the interests of fossil fuel producers.
As Drilled previously reported, a UN-affiliated organization helped seed the economic argument against acting on climate change, and a former Australian agricultural economist working within a tiny, innocuous government agency known as the Australian Bureau of Agricultural and Resource Economics (ABARE, pronounced “A-BEAR”) helped popularize and perfect the use of economic modelling to argue against strong climate policy. New documents obtained by Drilled from the Australian National Archives offer a glimpse into the Australian obsession with modelling, the dynamics that fed into making ABARE a boiler room for bad maths, and how the Australian diplomatic service sought to export the idea of modelling-driven climate policy to international climate negotiations in the hope it may protect the country’s coal and gas producers.
Rod Campbell, research director at The Australia Institute said the documents provide a remarkable window into the development of an idea, and are significant given the recent Australia election where “we’ve really seen modelling fail as a policy tool”, particularly when it was used by the conservative federal Coalition to justify the creation of a nuclear industry.
“What does modelling do? What does AI do? They flatten a complicated world to give easy answers,” Campbell said. “The world is complicated and any type of mathematical model simplifies that world and can be manipulated in ways to give the answers you want. The whole point of modelling is to make the complicated or subjective seem simple or objective and I think the same dynamics are at play.”
“What these documents show is how modelling has been a cover for business interests consistently right from the early days of neoliberalism, right up through to famous more recent examples like Adani’s claim its coal mine would bring 10,000 jobs.”
The genesis of Australia’s faith in the power of a spreadsheet begins in the primordial soup of the late eighties. Until that time, ABARE had been a dull little agency responsible for supplying statistics to farmers and stockmen. A bureaucratic consolidation forced by a Labor government seeking to present itself as “fiscally responsible” changed things when the agency found itself given the additional task of generating useful statistics for those working in an industry close to the heart of the Australian economy: resource extraction.
Australia had become the world’s biggest coal exporter in 1984 and over the course of the nineties it would grow to be one of the world’s biggest exporters of liquefied natural gas (LNG). According to ABARE’s 1988 corporate plan, the agency clearly understood its changing responsibilities, describing its role to “enhance the prosperity and economic growth of Australia through the conduct and dissemination in Australia and overseas of high quality, independent and objective economic research and analysis relevant to the agricultural, forestry, fishing, minerals and energy sectors of the economy.”
Funding cuts, however, would drive the agency’s actions over the next two decades. If knowledge is power, data is its foundation, but right as the public service was being pushed to do more with less, computers were being introduced to the working world. As the databases that contained the raw material needed by the agency began to grow in cost, ABARE faced the prospect it would be locked out of these systems and fall behind unless something was done.
"The Bureau's budget is coming under increasing pressure as a result of funding cuts initiated by both Government and the Department of Primary Industries and Energy,” an agency official wrote in an executive minute. “More and more the Bureau needs to look for ways in which it can supplement its resource base so that important research work is not jeopardised by externally imposed costs.”
The problem was so acute, the agency embarked on an extensive, forensic review—the file containing the results of these investigations runs to 250 pages—to study options for raising money to cover costs, and to figure out what rate they should be charging for their services. Their solution was twofold. First, the agency would take public information that was once free and begin to charge for these publications. Officially this was supposed to be against regulations, but government lawyers helped the agency find what it called a “loophole”.
“Whilst the Bureau cannot independently charge for its publications per se, there is no apparent impediment in recovering costs associated with the free distribution of our publications,” ABARE management wrote in an executive minute. “This loophole in the regulations has the potential to give the Bureau wide scope in earning revenue to offset publication distribution costs.”
The second line of attack was to pursue a closer relationship to industry. In its previous iteration, ABARE maintained a close relationship to industry bodies representing livestock companies, fishers, wool, sugar, cotton and dietary research groups. Internal documents within the agency even noted that a close working relationship between its teams and industry on special projects gave its staff motivation when working on large assignments. Working late nights in the office to strict deadlines, in close coordination with industry, made its employees feel like they were working for something, the agency reported, and it was keen to promote this going forward.
The “financial support” these industry associations and their member companies provided “offset some of the costs directly attributable to projects of special interest to specific industries.” All told, these contributions made up about 15 percent of ABARE’s budget in 1988, a figure that would more than double over the next few years as successive governments increased the amount of independent funding required by the agency. With the agency having taken over responsibility for supplying statistics to major oil, gas and coal producers – companies and industry associations with far deeper pockets than their counterparts in agriculture – it recognized an opportunity. Agriculture was giving way to minerals. Resource companies had more money than farmers and ABARE could charge a relatively fewer number of fossil fuel producers and their associations far more to keep up their regular output.
ABARE appointed economist Brian Fisher to head up the revamped agency and his role would be to follow the money. Fisher himself had started out as a consultant to ABARE, working initially in partnership with famed economist John Quiggin. When Fisher took over the agency, he ran it like a consultancy. Among his first tasks was to push through a restructure that involved a change in emphasis. Though they would still do work for agricultural producers, in a letter to the departmental secretary Fisher explained that “additional effort in the minerals and energy area will be achieved by a reduced number of sections in the agricultural area” with additional sections focussing on policy analysis.
Those early days would also bring Fisher into contact with at least two significant figures instrumental on climate change. One letter dated 1989 from Keith Orchison, then head of the Australian Petroleum Exploration Association (later renamed the Australian Petroleum Production and Exploration Association (APPEA), and today known as Australian Energy Producers (AEP)) complained bitterly the agency was not modelling his industry properly. Orchison went on to head the Electricity Supply Association of Australia, an organization that would similarly deploy modelling to influence the direction of climate policy. In that role, he would come to be known as the head of the “Greenhouse Mafia”—an in-house name once used by a small group of Australian political advisors, industry association heads, company officials and bureaucrats to describe themselves as they lobbied to undermine any government response to climate change and keep Australia from signing the Kyoto Protocol.
Meanwhile, an organizational chart from the same year lists Barry Jones as a section manager with ABARE’s petroleum economics team. A creature of the Australian public service, Jones would later take over as head of APPEA in 1996, using his time with ABARE as a selling point. In that role he would leverage his deep experience with the public service on climate change as part of the Greenhouse Mafia to coordinate the political activities of the industry alliance it represented in Canberra.
Over the ensuing years, Fisher courted APPEA, giving presentations at their conferences and contributing as a co-author to The APPEA Journal. These relationships would prove useful as Fisher’s ABARE became increasingly focussed on developing its modelling muscle. A group of academics and consultants, including economists with the Melbourne University Institute of Applied Economics and Social Research working with an economic model called ORANI, were first hired in 1989 by resource company CRA to assess the implications of a plan to reduce emissions 20% by 2005. CRA, later renamed Rio Tinto in 1997 after a merger, was a producer of coal and uranium. The company would maintain an interest in ABARE’s work in this area under Fisher.
Internal documents show ABARE’s teams weighing the cost and effectiveness of available databases and early models – all of which were unsatisfying. Either they were too limited, too expensive, or didn’t have the capability the agency wanted. By 1993, ABARE was at work developing its own model, MEGABARE. Unlike other models, MEGABARE’s advantage was that it married trade flows with a model of the national economy, where others did not. To fund its creation, Fisher pitched the Business Council of Australia, the Australian Coal Association and APPEA membership for financial support.
Like AI developers in the present, Fisher’s genius was in marketing. He understood how to speak to his audience in a language they understood. In one paper he co-authored and presented to an APPEA national conference, Fisher pointed directly to climate change as an example of how a lack of information led to uncertainty that made it hard to act.
“While it may be possible to determine the extent of greenhouse gas emissions resulting from a given level and mix of economic activity, the effect of those emissions on the level and mix of future economic activity is difficult to ascertain,” they said.
Modelling, he suggested, could change that. To demonstrate, the following year, Barry Jones, working in his capacity within ABARE, co-published an article in the journal Energy Policy with the innocuous title, “Reducing Australian energy sector greenhouse gas emissions”. Jones, as lead author in the three-person ABARE-team, used economic models of the Australian energy sector to estimate the cost of climate action. His conclusion would become a well-known objection in Australia to suggestions that the country act on climate change: it would cost far too much.
“The changes to the electricity system alone […] would result in a reduction in annual growth in domestic product of about 0.3%, and a somewhat larger reduction in growth in real wages, accompanied by a substantial devaluation of the Australian dollar,” he and his co-authors wrote.
“Detailed analysis indicates that the target can be met by the energy sector, but only with substantial change to the structure of the sector, achieved at considerable cost,” they warned.
ABARE would cite this same article when it was contracted in August 1994 as a consultant by the Department of Foreign Affairs and Trade (DFAT) to run a more ambitious study on the implications of addressing climate for Australia using MEGABARE. The file containing these documents from the controversial project runs to 258 pages that, according to climate transitions scholar Marc Hudson, offer a rare glimpse inside the mechanics of the Australian state at a key moment in its history.
“DFAT’s concern would have been that they will rock up in Berlin in 1995 and be hammered by demands that are against the short, medium and long-term interests of the oil, gas and coal exporters,” Hudson says. “They absolutely associate the fossil fuel producers’ interests with Australia. In the same way that what’s good for GE is good for America, what’s good for BHP and Rio Tinto is good for Australia.”
“DFAT are planning for the worst; hoping for their best. They’re thinking they want to go out to international meetings with these numbers, arguing that Australia would be unfairly hit, that Australia needs exemptions – the stuff John Howard did in 1997, these documents suggest DFAT was already planning for that.”
Drawing up the project brief, DFAT instructed ABARE to use its model to “provide input into determining an appropriate Australian negotiating position on international agreements on policies to address climate change.” DFAT also made clear the study “would involve ascertaining the likely effect of proposed international commitments and outlining alternative policies that might suit Australia” owing to concerns about the financial cost of addressing climate change to Australia's fossil-fuel dependent economy.
“The relatively high cost, by OECD standards, of reducing greenhouse gas emissions in Australia means that, if any stringent international targets were made binding, it would be likely to lead to significantly higher adjustment costs in Australia relative to most other countries,” the department said.
That cost, however, only focussed on one side of the ledger. The cost of rising insurance premiums from rolling natural disasters amplified by climate change or the harm to agricultural production from a breakdown in predictable seasons was not considered. An appendix attached to the brief shows that the only real consideration was the cost of regulation to Australian fossil fuel production and export. Bar graphs showing how bad Australian oil, gas and coal producers, and those of major trading partners, would be hit under various climate policies suggested the number would be in double digits, but less under certain, preferred policy options.
Briefing ABARE, DFAT asked the agency a question one might ask ChatGPT today: to advise on “possible coalitions” Australia might join while taking part in international climate negotiations.
“MEGABARE will be used to assess the economic advantages to Australia of forming a number of potential coalitions,” it said. “Implications will be drawn for Australia's negotiating strategy. Potential coalitions include Australia as a member of a coalition with: i) individual Annex 1 or non-Annex 1 countries; ii) with APEC members iii) with a group of major developing countries (iv) a global grouping; and (v) fossil fuel producers."
Both agencies were explicit about how this was a play for influence on the direction of Australian climate policy: "It is important that the study focuses on informing and influencing the position of policy makers and increasing awareness generally of the issues involved in climate change”, the brief reads, even as it acknowledges the critical part was presenting results to the public.
“A large part of the process of informing the debate within Australia will depend on presentation of the arguments behind the model results as well as the results themselves,” the project brief read.
To achieve this, ABARE and DFAT worked to sideline the Department of Environment, Sport and Territories (DEST) from the project. DEST officers at the time were old school. They were, largely, straight shooters who sought to operate with the objective detachment and even-handedness demanded of the Australian public service. They viewed the creation of MEGABARE as having merit and even contributed funds to its development, but were also aware that the science of climate change, and the economic and social cost of inaction was of little interest to their counterparts in the Department of Primary Industry and Energy.
A “steering group” was formed that included other government agencies, but ABARE made clear they would have no influence over the process or final presentation of the work. Meetings of the steering group took the form of briefings where a representative from ABARE would give a presentation about MEGABARE’s development. Letters between the departments show DEST officials were suspicious from the outset. One early letter offered a cool response to the study proposal, saying DEST is “interested to know the timing of the project given the teething problems ABARE has had in getting the MEGABARE model running”.
“What you’ve got is an intra-state turf war, with people calling each other arseholes in polite language,” Hudson says. “DEST is arguing that we can do emissions reductions much more cheaply than the Business Council of Australia is saying, but what ABARE and DFAT are saying is that the issue is ours.”
“Consultants know politicians are suckers for numbers and bar graphs and charts. ABARE is trying to prove to everyone that it is a responsible and adult body, and bar graphs help that.
“Meanwhile, by using ABARE as consultants to draw up modelling in this way, the tactics that DFAT are using here are the same tactics that had been used with great success domestically, and they were hoping to repeat the trick, globally,” Hudson says.
One letter from DFAT in May 1995, appears to make that intent clear. At the time, Australia was the exception for its interest in using modelling to shape how it approached international climate negotiations. While providing feedback on early drafts of the report, DFAT directed ABARE to use its model to promote “an optimal strategy” that would benefit Australia’s economic interests in future international climate negotiations – even if it acknowledged success was unlikely.
“Refer to an attachment that discusses this optimal strategy using MEGABRE and briefly set down the key points,” one instruction began. “But this approach presents difficulties from the negotiating perspective: would parties to the negotiations be willing to accept model generated outcomes? The answer may well be no and, even if that were not the case, there would be a long and complex negotiation about methodologies.”
With Brian Fisher working at the time as a lead author on an IPCC working group dealing with the use of economic modelling in developing climate policy, control over ABARE’s work in this area would be increasingly devolved to Vivek Tulpule – who today works for Rio Tinto in Singapore as head of economics. Despite the agency's best efforts, it failed to meet its deadline – though this did not stop DFAT from promoting the study among Australian diplomatic outposts. Even before the study had been printed, DFAT sent a cablegram to Australian diplomatic missions from Paris to Port Moresby, to gauge interest in a “major study on possible least cost approaches to limit global greenhouse gas emissions in the period 2000-2020.”
“The study uses MEGABARE, one of the most sophisticated general equilibrium models so far developed for examining environmental policy issues,”the cablegram read.
The response was overwhelming. In reply, DFAT received requests for a total of 1246 copies across the 29 missions. Beijing wanted 60, New Delhi and Copenhagen wanted 50 each – Japan asked for 500 copies prompting a hand-written note on the file asking: “are they serious?” And another, commenting: “Looks like a real hit!”
ABARE and DFAT’s glory, however, would be short lived when an early draft of ABARE’s report leaked. Newspaper clippings collected by ABARE include an article published in the Sydney Morning Herald in August 1995 with the headline: “Coal lobby ices our eco-credibility” and “Secret strategy undermines greenhouse fight”. The next day ABARE and DFAT officials were called in to explain themselves in an emergency meeting. Hand-written notes suggest the newspaper reports touched on “sensitive policy issues” and the Department of Prime Minister and Cabinet worried about the public reaction. Departmental officials complained of a “litany of false assertions” in the reporting that “deliberately set out to create disputes (polemic)" – chief among them the claim that the study had been paid for by the coal industry.
ABARE’s executive director sought to defend the project in a minute addressed to the Minister for Primary Industries and Energy where he stressed the study was “not funded by the coal industry”. But even as the agency emphasised the study was “jointly funded by DFAT and ABARE”, it conceded the “construction of MEGABARE” was “funded by a broad range of industry groups and government departments.”
“Contributors include DEST, BHP and CRA (the latter two through the Business Council of Australia), DFAT, BIE, NSW Coal Association and DPIE,” read the brief from ABARE’s executive director, titled “Negative publicity regarding ABARE’s work on greenhouse issues.”
This, however, was not a complete list. An investigation by the Commonwealth Ombudsman in 1998 would report that the Australian Coal Association had been the first to fund the creation of MEGABARE in 1993, before being joined in later years by organisations such as Australian Aluminium Council, Exxon, Mobil and Texaco.
Even as the agency’s heads were writing to newspaper editors “in respect to inaccuracies in the articles”, the diplomatic damage had been done. A cablegram from Australia’s diplomatic station at the United Nations in New York sent a few weeks later summarised the latest edition of environmental magazine Eco that included an article about the ABARE climate study. Borrowing from domestic Australian reporting, it informed its audience of international environmental groups the Australian government “aligns with the coal industry” and will release “a highly controversial economic modelling study, partly funded by the coal industry; which claims that developing countries will suffer if developed countries reduce their emissions”. The Eco’s reporting had been “commented on informally and in passing by various other countries”, the diplomats said, and had “drawn some interest from representatives of business organisations.”
Even if the shine had been taken off its work, the ideas ABARE’s numbers advanced would dominate the Australian discussion about climate change. For the next decade, conservative political leaders would cite ABARE numbers to argue any effort to address climate change would be too expensive, disastrous for the economy and that it was best to do nothing.
Today, the dynamics may be different. ChatGPT may be more autonomous and operate more as a black box. The precise ideology may diverge at a certain point, but Dr Marc Hudson says, at their core AI and economic models perform a similar function: to advance the interests of a few over the many.
“It targets this endless faith in numbers and computers that is one of the human species' chronic design flaws,” he says. “Any conversation about whether historical, or present, or AI or old school, has to include the fact that models are produced for a political purpose. Either to support a policy or combat a policy. No one does it for fun.”
Amy Westervelt and Dr Marc Hudson contributed documents to this reporting. Drilled contacted Dr Brian Fisher through his publicly listed email address at BAE Economics with detailed questions. He did not respond.
