
Stop data centre construction in our suburbs
AI is being developed by corporations in competition for profit and military advantage, with governments fast-tracking their approval every step of the way. These companies train commercial models on stolen creative and personal work without consent or payment, and are dumping polluting, power- and water-hungry data centres on working-class communities with next to no consultation. Bosses are already using AI as an excuse to cut jobs and intensify work while pocketing the savings, and governments are quietly handing life-affecting decisions to unaccountable algorithms. Victorian Socialists believe AI should be put under democratic public control, not left to the market, the war machine, or the surveillance state.
What we think
- Advanced AI is too important to be controlled by corporations competing for profit and military advantage. The most capable systems should be treated as public infrastructure.
- Current laws don't stop tech giants stealing people's creative and personal work to train commercial models, or dumping hyperscale data centres on working-class suburbs with no real say for residents.
- AI-driven productivity gains should be directed to reduced hours and increased pay for workers.
- Automated decision-making about welfare, healthcare, policing or bail has already proven disastrous. Robodebt showed what happens when algorithms replace human judgement and accountability and an episode like this should never happen again.
- We recognise that unaccountable, dangerously capable systems, and companies whose business models are built on surveillance and repression, pose real risks that governments should act on before they cause any further harm to people or the planet.
We'll fight to
- Put major AI infrastructure and computing capacity under public ownership, and create a Victorian Public AI and Computing Institute for public-interest research and computing capacity.
- Require registration, independent safety testing, incident reporting and government power to halt the training or deployment of frontier AI models, with developers legally liable for harms, including a moratorium on models that can't demonstrate they're safe.
- Require full disclosure of AI training data sources, and consent, credit and payment for any copyrighted, creative or intellectual work used to train commercial models.
- Introduce GDPR-style protections making personal data the property of the individual with the right to access, list and delete it on request, mandatory Australian data storage, and criminal penalties for unauthorised sharing.
- Ban autonomous weapons, predictive policing, facial-recognition surveillance, and AI-driven decisions on welfare, healthcare, immigration, bail or sentencing.
- Ban Palantir from operating in Victoria or holding government contracts.
- Give workers and their unions the right to veto workplace AI that cuts jobs, pay or hours or increases surveillance
- Guarantee paid retraining and income protection for any worker made redundant because of AI.
- Mandate community consultation, environmental impact assessments, on-site renewable power and high commercial water and electricity rates for all AI data centres, ban public subsidies for them, and ban new hyperscale data centres in residential areas.
- Pause all data centre construction until the above conditions are met.
- Rule out any public bailout of corporations if the AI bubble bursts
- Fund and introduce AI critical literacy in public schools, TAFEs and universities.