Artificial Intelligence and Sustainability: When it Takes Energy to Save Energy
David Blanch, director of product management, ESG and Environmental Intelligence at IBM, provides commentary on “The State of Sustainability Readiness Report”
Is artificial intelligence (AI) the key to meeting sustainability goals or is it simply the latest shiny object? IBM’s new “The State of Sustainability Readiness Report” helps answer this question by looking at the current trends surrounding technology and sustainability.
Perhaps one of the most surprising findings in the report is that nine out of 10 business leaders surveyed “believe AI will positively influence achieving their sustainability.” That such a large majority are looking to AI to help improve sustainability brings up numerous questions about how the technology might be implemented, while some organizations pioneer “carbon-conscious” uses of AI.
The report also examines some of the challenges organizations face related to the amount of energy AI consumes, and makes recommendations for organizations exploring the use of AI to further sustainability goals. David Blanch, director of product management, ESG and Environmental Intelligence at IBM, spoke with TechChannel and offered additional thoughts about the report’s findings.
Sustainability Is a Hard Problem
Business leaders don’t just think that AI can improve their sustainability efforts; they are also investing in it. The IBM report reveals that 88% plan to increase investments in IT with the goal of furthering their sustainability efforts.
Businesses have good reasons for working to improve sustainability:
- As the public becomes more aware of the need to reduce carbon emissions, the brand reputation of any company that doesn’t explicitly work to improve sustainability is in jeopardy.
- It’s just good business to become more efficient as energy costs rise.
- Reliability and resilience must be considerations in a world where climate change is leading to more natural disasters that can disrupt services, such as floods and wildfires.
- Regulatory pressures motivate even those organizations that may be reluctant to embrace sustainability.
“It’s simple to say, ‘I want to reduce my emissions, or I want to become net zero by a set target date compared to a set baseline year.’ But it’s a hard problem to solve,” says Blanch. Sustainability isn’t one thing; it requires participation from every department and team, and even vendors, suppliers and partners. When it’s a proper goal, sustainability becomes a performance indicator across the organization.
The Work That Must Come First
Since tackling emissions and improving sustainability is complicated, a tool that looks like it can serve as a singular solution may be appealing, but most organizations have work to do first.
“There are a number of things that organizations have to do in order to get their house ready before they can start using technologies like AI,” says Blanch.
One case study on sustainability efforts incorporating AI can be found with the Downer Group, which builds and supports the operation of trains used for both passengers and cargo in New Zealand and Australia. The company’s goal is to reach net zero greenhouse gas emissions by 2050, and it’s using technology to make important progress.
Downer is working with IBM Consulting to develop and enhance a platform called TrainDNA that supports predictive maintenance of their trains in Australia. From a customer perspective, that means they are more reliable, and from an efficiency standpoint, more trains can be maintained at the same time because the maintenance is planned. The company has realized a 20% improvement in efficiency.
Before all of that could be implemented, though, Downer needed insight into what was actually happening. With the TrainDNA platform, the company has enhanced observability. They can see where all the trains are, what their timetables are and much more. The platform creates a single place for all of that data that was previously available only separately.
Everything comes back to data. Over the last decade or two, data has become increasingly valuable in almost any context you care to name, and the technologies used to run businesses of all types require massive amounts of it. The problem is that all of that information is stored in different spreadsheets that are only available to certain people, or in video, handwritten notes, or scanned documents.
Beyond the fact that data can be stored in a vast array of formats, from paper to images to spreadsheets, it’s also often locked away within particular departments across organizations, with each one following rules and regulations that may not apply to others. This was true for Downer, and before AI could improve reliability and efficiency, all of the different inputs from maintenance centers, stops, ticket sales and more had to be brought together and synthesized so that it could be analyzed.
“It’s almost of a question of how to get data to make sense, and use AI to streamline the process, and gain visibility into my current footprint, and then devise some strategies,” Blanch says, noting that the phrase “deploying AI” is much more involved than it might seem.
The Balancing Act: AI and Energy Consumption
AI can help organizations reach their sustainability goals. AI consumes large amounts of energy.
Both statements are true, and making sure that sustainability goals are met even while using a technology that requires the use of vast amounts of energy requires careful navigation. Blanch points out that all activities consume energy, but AI also “has a lot of potential to help streamline, automate and help organizations spot things that otherwise might take more resources to accomplish.”
Since meeting sustainability goals involves the entire organization, its suppliers, vendors and partners all collaborating and working together, the opportunities for becoming more efficient are huge, and “as organizations get more efficient, and deploy automation technologies, they can dynamically scale the resources they need up and down as needed,” says Blanch. Some of the practices the IBM report recommends include:
- Using foundation models
- Optimizing data processing locations
- Using energy-efficient processors
- Taking advantage of open-source collaborations
One positive aspect of the massive popularity of AI is that it’s galvanizing companies to look for ways to be more energy efficient throughout their value chains, resulting in improvements in sustainability that would likely have gone unrecognized otherwise.
The Maturity Curve Matters
Another interesting finding of IBM’s new report is that business leaders are more enthusiastic about the potential of AI than the people who report to them. The report found that 67% of C-suite executives believe their organization is taking a proactive stance when it comes to climate resilience, but only 56% of vice presidents and directors believed the same. Similarly, C-suite leaders are at the upper end of a 12% gap between the two groups regarding their belief that AI will help their organizations reach sustainability goals.
Blanch suggests that these gaps have to do with companies’ maturity curves and the sophistication of their goal setting, along with several external drivers such as how a given industry is regulated, the approach to climate and sustainability in a specific geography or industry, and even company culture. “As companies are maturing, the data problem is just becoming more complex as the companies are,” he says.
Blanch advises that before AI can deliver results around sustainability, companies must make sure they have deep visibility into things like application and asset performance, so that any issues related to the reliability of their systems can be identified.
The next step is related to resource allocation, and deep consideration of how resources are managed. It is important to consider strategic planning for budgets and physical assets when introducing, optimizing and prioritizing sustainability. Taking those steps before introducing AI makes sustainability a core concept that penetrates all the various internal organizations rather than something that is managed centrally.
By following that approach, the same business leaders who predict AI will be a net positive for sustainability are also empowered to help make that vision a reality.