A new study has revealed that AI may be a net negative for the environment rather than a net positive when considering the full energy sector.

Researchers compared the potential for AI to increase clean power generation alongside its potential to produce coal, oil and gas – it found that across 64 scenarios, net carbon pollution would rise 0.47-1.8 gigatonnes a year, equating to 1-5% of the energy sector’s overall annual emissions.

The research took the same principles AI developers have touted as far as renewable energy production – optimisation of the energy grid, reducing downtime – to oil and gas procurement.

While AI may make the renewable energy industry more efficient, it can also do the same for oil and gas – it can help identify new sources of fossil fuels and speed up their recovery, leading to an increase in fossil fuels and carbon emissions.

If AI is adopted by each industry at the same pace, the research paper posits, the productivity rates of the renewable industry found have to outpace the fossil fuel industry at least four fold in order for emissions to break even.

The paper notes that relatively, the renewables industry is largely in its infancy compared to the fossil fuels industry, which is adopting AI already.

Estimates from the International Energy Agency says that AI can increase technically recoverable fossil fuel reserves by 5%, and can cut the cost of deepwater offshore drilling by 10%.

Recommended reading

• Can AI De‑Risk Breakneck Data Centre Builds?
• Will AI Energy Demand Undermind Climate Gains?
• Edinburgh Council Rejects Data Centre Proposal
• How Badly Will AI Data Centres Be Impacted By Power Shortages?
• Are Traditional Data Centres Fit For Purpose?

Meanwhile, renewable energy efforts are still largely in the pilot or research phase, without longstanding international infrastructure and funding to catch up with the fossil fuel industry.

While AI has largely been touted as helping to make energy use more efficient, the paper makes a point of applying this same efficiency and productivity principle to the fossil fuel industry itself.

“Absent policy steering, AI’s modeled effects increase the carbon intensity of the global economy and reinforce fossil fuel incumbency—outcomes that current analytical and governance frameworks do not fully capture,” the research paper, written by Will Alpine, Nathan Geldner, Holly Alpine, and Maksym Chepeliev, said.