On January 15, the Australian Council of Trade Unions (ACTU) and Microsoft announced a new “framework agreement” on the implementation of artificial intelligence (AI) in workplaces across the country.

The ACTU claimed the agreement “sets a new benchmark for workers’ rights in the tech sector.” In reality, it is a thoroughly pro‑corporate arrangement designed to legitimise the widespread replacement of jobs with AI, and a pledge from Australia’s top union bureaucrats to suppress workers’ opposition to the rollout.

The ACTU and the bureaucracy are completely committed to the logic of the capitalist system, under which every advance in technology and productivity is used to slash pay and conditions. It is intensely hostile to any struggle against the profit system, the only means by which the potential of technological advances, such as AI, can be used to benefit workers, not the corporations.

The ACTU-Microsoft deal has been hailed by the Labor government, business groups and the corporate media as a model for how unions and major corporations can jointly impose a massive expansion of AI throughout industry.

It builds on a separate Memorandum of Understanding already signed between Microsoft and three ACTU affiliates—the Australian Services Union, the Shop, Distributive and Allied Employees Association and Professionals Australia. The earlier document enshrined the rights of these unions to cover Microsoft workers, thereby installing them as the official overseers of the “AI transition.”

The “framework agreement” is explicitly not legally binding and imposes no limits on the company’s capacity to eliminate jobs.

Instead, it outlines three core objectives:

Information Sharing, under which Microsoft will run training sessions for union leaders and staff through the ACTU institute. In other words, the bureaucracy will be schooled by corporate experts in how to sell the AI transition to workers as inevitable and even beneficial, and in how to divert anger over job losses and intensification of labour into illusions in “upskilling” and “lifelong learning.”

Worker Voice in Technology Development, a euphemism for “consultation” processes, carefully stage-managed by the union bureaucracy and management, aimed at giving sweeping restructuring measures a false imprimatur of worker approval and involvement. Under the pretext of “incorporating worker perspectives into the design, development and deployment of AI systems,” workers will be dragooned into architecting the destruction of their own jobs or those of their colleagues.

Moreover, such mechanisms do not alter the basic fact that the underlying strategic decisions—where and when AI systems will be deployed, what roles will be automated, and which sites will be downsized or closed—remain entirely in the hands of Microsoft’s executives and shareholders.

Collaboration on Public Policy and Skills, a commitment from the ACTU to join Microsoft in lobbying to ensure that any regulation on the use of AI in the workplace is compatible with corporate profit demands. The framework is proudly linked to the Labor government’s union-backed National AI Plan, announced last month, which rejects legal constraints on AI deployment in favour of voluntary principles drafted in close collaboration with industry.

The ACTU is promoting its accord with Microsoft amid a global wave of AI‑linked restructuring, in which the company itself is playing a leading role. Last year, the company slashed more than 15,000 jobs worldwide in multiple rounds, under the vague pretexts of “efficiency,” “business priorities” and workers’ alleged “poor performance.” While Microsoft did not openly admit that these cuts were related to the expansion of AI, industry analysts generally take it as a given. The company invested some $80 billion in AI in 2024–25 and its executives were already bragging last year that almost one-third of software coding at the firm was done by AI.

Demonstrating the company’s attitude, an executive from its games division posted on X after one round of cuts, “I’ve been experimenting with ways to use LLM AI tools (like ChatGPT or Copilot) to help reduce the emotional and cognitive load that comes with job loss,” suggesting that sacked workers could simply seek solace from the same technology that replaced them.

Last year’s global cuts at Microsoft followed its elimination of 10,000 jobs in 2023 and an estimated 4,500 in 2024.

All of this has taken place since Microsoft in mid-2022 brokered a “historic neutrality agreement” with the Communications Workers of America (CWA) and established “labor principles” for collaboration with unions more broadly. This was expanded and formalised with the December 2023 announcement that Microsoft had formed a “new AI alliance” with the American Federation of Labor and Congress of Industrial Organizations (AFL-CIO).

In other words, the US deals that were the direct precursors to the ACTU agreement have provided the mechanism through which Microsoft has slashed almost 30,000 jobs in three years.

This should be a stark warning to all Australian workers, not just those employed by Microsoft or in the technology sector, of the carnage to be imposed across broad sections of the working class through the joint action of the unions, big business and government.

Federal Labor MP Andrew Charlton praised the Microsoft-ACTU agreement as a “positive step entirely in line with the National AI Plan, which called for a “consultative approach to AI adoption in the workplace” and vowed to “bring together government, unions and business.” 

A mere two weeks after the National AI Plan was announced, complete with promises to protect workplace rights and “enable workers’ talents, not replace them,” Treasurer Jim Chalmers announced that the government would slash $6.8 billion over four years from the budget by gutting the public service, specifically vowing to “unlock the full potential of AI in public service delivery.”

To date, Australian corporations have proceeded somewhat cautiously, invoking various pretexts to explain the culling of jobs while trying to downplay the relevance of their growing investment in and implementation of AI.

Most notoriously, the Commonwealth Bank of Australia, which in July was the first employer in the country to openly announce the replacement of workers with AI, was forced to retract the announcement the following month, facing a public backlash and a hearing in the Fair Work Commission.

With the National AI Plan and the ACTU-Microsoft agreement, Labor and the union apparatus are sending a clear signal to big business that such caution is no longer required. The union bureaucracy will work closely with corporations to impose their AI plans and Labor governments will ensure the absence of any regulatory hurdles.

While the technology is new, the role of the union apparatus and Labor is not. Their commitment to the pro-business AI transition today is rooted in the ACTU-Labor Accords of the 1980s and 1990s, through which the always nationalist and pro-capitalist program of the unions was taken to its logical conclusion. These increasingly corporatist entities were charged with enforcing sweeping cuts to jobs, wages and conditions, under the guise of making Australian capitalism “internationally competitive.”

Serving as an industrial police force, the unions have over the past four decades presided over the destruction of vast swathes of jobs and even entire industries, suppressing workers’ opposition through the same empty promises as are contained in the new AI agreement: consultation, retraining, upskilling and redeployment. The reality for workers has been the gutting of whole towns and suburbs through long-term unemployment and a massive downturn in wages.

To defend jobs, wages and conditions, workers at Microsoft, in the tech sector more broadly and across all industries must draw the necessary conclusions. The struggle against AI‑driven restructuring cannot be entrusted to organisations that function as partners of management and the government. It requires the formation of rank‑and‑file committees, democratically controlled by workers themselves, independent of the unions and the Labor Party.

Such bodies would begin from the opposite standpoint to that of the ACTU‑Microsoft pact. The introduction of new technologies, including AI, must be based on the needs of workers and society as a whole, not what delivers the greatest profits for shareholders.

The issue confronting workers in the AI transition is not the technology itself, but who owns and controls it. AI could be used to eliminate onerous tasks and shorten the work week without any reduction in output—and therefore pay—giving workers more time for family, recreation and cultural activity. But in the hands of the ruling class, AI, as has been the case with every technological breakthrough for over a century, will be used to intensify the exploitation of workers and increase profits.

This poses the need for a political fight against capitalism, under which the progressive content of every technological advance is subordinated to the interests of the financial and corporate elite. The alternative is a socialist program: the expropriation of the tech giants, banks and major corporations, their transformation into publicly owned utilities under democratic workers’ control, and the reorganisation of economic life on the basis of human need.

While the ruling class is using the development of AI to intensify the assault on workers’ rights and conditions, they cannot eliminate the fundamentally progressive potential of the technology itself. The working class must make use of this scientific advance to take forward the fight to end capitalism. To arm workers for this struggle, the International Committee of the Fourth International has developed Socialism AI, a chatbot that provides workers with access to a vast trove of Marxist theory, history and strategy to help build the revolutionary socialist movement that is required to defend jobs, wages and conditions through the abolition of the capitalist system.