There is little doubt that the climate crisis is a challenge like no other, with each year briging higher average temperatures and more extreme weather events. AI-optimists look to AI to help tackle the climate challenges we face, but others note that AI may be part of the problem.
We asked some industry experts for their insights into AI, and how it might help us towards sustainability.
Working sustainably
TrusTrace works with the supply chain to make sustainability easy for companies, collecting the relevant data to inform their production decisions. “TrusTrace customers use this data to reduce products’ sustainability footprint across various environmental, social, and governance parameters,” Shameek Ghosh, co-founder and CEO, explained. However, even in a function that might seem straightforward, AI is being used.
“The use of AI has helped us digitise the value chain faster,” Ghosh says. “The impact of using AI has been the ability to collect different kinds of data and digitise it at a rate three times faster than without AI technology.” As well as helping with data analysis and collection, Ghosh says that AI also helped them expand their product coverage, helping them meet the challenge of supply chain sustainability. In essence, AI helps them collect more information, about more products, faster than before.
For Ghosh, the adoption of AI was part of his approach. “A transformative CEO is always looking out for the disruptive forces within their industry, and doesn’t take their company’s position for granted,” he said. “The person is quick to acknowledge new changes and adapts to those changes — they know how to drive change within the organisation and take their team with them in the process.”
Offering sustainable services
As well as company leadership, some note that customer demand is driving improved sustainability. Alina Arnelle, Chief Sustainability Officer at BeCause, a company making sustainability management easy for the tourism industry, highlights a McKinsey survey that found 78% of people consider sustainability important to them. And that affects the choices they make.
“Hotels with eco-certifications convert four times more guests compared to ones that don’t have sustainability certifications,” Arnelle told us. “When taking into account guests’ willingness to pay more for green hotels, we estimate that poor sustainability management results in $13 billion in missed revenue.”
While the environmental cost of flying has long been highlighted, the travel and tourism sector generally has a long way to go. “Hotels, in particular, contribute to environmental degradation in various ways, including energy consumption, water usage, and waste generation,” says Arnelle.
And AI has been used to help improve sustainability. “We’ve seen an increase in the adoption and integration of AI technologies as they become more accessible across industries,” Arnelle told us. “From efficient data collection to the augmentation of sustainability strategies, these solutions are designed for use by staff at the property level as well as the higher-level decision-makers tasked with overseeing sustainability initiatives across the group.”
Making AI itself sustainable
But he AI industry itself is receiving criticism for its environmental impact.
“AI presents many opportunities to reduce climate change or reduce the impact of climate change,” Phil Burr, the Head of Product at Lumai, an Oxford University spin-out creating 3D optical computing for AI, told TFN. However, he added, AI comes at a cost.
“It’s a challenge because current AI systems consume an incredible amount of power,” Burr explained. “One study has even estimated that responding to 2,000 AI text generation requests uses approximately the same energy as charging a mobile phone.”
Most of us charge our phones every day, but at scale, it’s a problem. “When you multiply this by the number of queries, the total energy suddenly becomes significant,” Burr continues, “it’s estimated that ChatGPT has 60 million website visits every day.” When you consider that each of those visits will result in multiple queries, then factor in the other AIs using power, it’s easy to see that the power demand for AI is significant.
And that demand is only growing. “It has been reported that data-centres will double their electricity consumption in just two years; equivalent to as much energy as the whole of Japan,” says Burr. But even if AI companies want to burnish their green credentials, they still have a significant impact, even if they use green energy, that simply stops someone else from going green.
Lumai is creating new, more efficient ways of addressing this specific challenge. “A lot of innovation will come from hardware. New technologies, such as Lumai’s, where light is used to perform the AI computation, can do this many times faster using a fraction of the energy compared to today’s semiconductor solutions.”
Barring unexpected regulation, AI is here to stay. It will, undoubtedly, bring benefits, but if AI is to avoid being part of the sustainability problem innovators and startups will need to address their use. Burr expects to see that happen. “Demand for AI will continue to rocket,” he said. “So, whilst we don’t expect the drive for more and more AI processing to abate for the foreseeable future, we do expect a greater focus on energy efficiency.”