3 ways regulation can prevent deepfake greenwashing

More than half of company leaders admitted to greenwashing in a 2022 survey of nearly 1,500 executives. Meanwhile, an EU Commission study shows 40% of green claims have no supporting evidence.

Greenwashing erodes public trust in climate action and the transition to net zero. It also makes it difficult to inspire and encourage other companies and leaders in the public and private sectors to join the fight against climate change.

And with traditional greenwashing being so common, there’s a risk it could increase with the advent of deepfake technology, which provides the capability to fabricate convincing visual and auditory content. The total number of deepfake videos online had already risen to about 95,820 in 2023 – 500,000 deepfake videos and audio content on social media platforms alone – a 550% increase from 2019. Last year it was estimated that 8 million deepfake videos will have been shared online by 2025.

Deepfakes could be used to falsely demonstrate commitment to renewable energy, carbon capture and sustainable development, for example. In March 2023, a think tank was accused of using a doctored image of a dead whale near wind turbines for a newsletter arguing against off-shore wind projects. A spokesperson for the organization, the Texas Public Policy Foundation, called it “stupid” to refer to the image as “doctored”, arguing “the graphic was made to advance a point made in an article”, according to the Guardian.

That same year, a doctored video of environmental activist Greta Thunberg circulated online in which she praised vegan grenades and sustainable tanks. The original video was a BBC interview where she was talking about climate anxiety.

These are just a few instances that show the emerging threat deepfakes could pose to climate action and sustainability. And this emerging threat of deepfake greenwashing underscores an urgent need for regulatory safeguards against such deceptive practices.

The European Union AI Act stipulates that all AI-generated content must be disclosed as such. The Act regulates Generative AI (GenAI) by mandating that videos, audio files and images should be clearly labelled. This will help those operating in Europe to distinguish AI-enabled greenwashing from genuine climate action.

The US currently has no federal legislation on AI, but several states including Colorado, Tennessee and California have established or begun to enact AI legislation, with most due to be in force from 2026. Notably, the Utah Artificial Intelligence Policy Act requires individuals and entities to disclose the use of GenAI when dealing with consumers.

China’s Interim Measures for the Management of Generative AI Services (2023) regulates deepfakes and synthetic content, which would limit the threat of GenAI-induced greenwashing. This is one of the few AI policies that regulates the use of GenAI, particularly deepfakes.

And while there is no African Union AI treaty yet, the African Union Development Agency (AUDA-NEPAD) has established a Continental AI Strategy as a framework for ethical and responsible development of AI in Africa. This strategy is good, but regulations are more effective because they are permanent and entrenched, and so more difficult to reverse. The African Union has been unable to establish a formal AI regulation due to a lack of unity and cooperation from member countries.

Mauritius and Egypt released National AI strategies in 2018 and 2025 but, to date, they have no specific rules that cover deepfakes or climate action.

With deepfake technology developing faster than regulations, there is the risk of a great rise in greenwashing. Regulation must play a proactive role in preventing the dangers that deepfakes may pose to the fight against climate change. Here are three ways governments can help prevent deepfake greenwashing:

Regions and continents like Africa must adopt AI treaties that provide standards and frameworks within which GenAI can be ethically and legally deployed. In particular, these AI treaties should include safeguards against the use of GenAI in climate action disclosures and include punitive measures for deepfake greenwashing.

Countries that already have AI policies in place must go further to establish AI legislation that can cater to the emerging threats that deepfakes could pose to global climate action.

As well as enacting legislation on the regulation of AI, governments should collaborate with industry and sector regulators to incorporate workable AI provisions into their respective legal frameworks.

Strong cooperation will help to establish a harmonious legal framework that covers the use of GenAI in various sectors.

AI is still evolving. Laws and regulations should not just focus on the present position of GenAI, governments should also incorporate foresight and future-based thinking into the drafting of regulations and laws.

This will help to protect both current and future commitments to climate change mitigation and adaptation. Legislation must be strategic and for the long term to prevent deepfakes – and whatever AI developments come next – from obscuring genuine local and global climate action.

Greenwashing powered by GenAI will make it even more difficult to measure the extent of progress being made to combat climate change. If we can’t differentiate the true from the fake efforts at climate action, it may become increasingly difficult to measure climate action achievments.

Robust AI regulations and legislation must govern the ethical implementation of GenAI, especially in the climate action space.