The disruptive upswing in expectations and concerns surrounding AI continues at lightning speed, as businesses come closer to understanding the rapidly changing technology and its potential functions within their business. Recently, advanced AI platforms such as generative AI and LLMs (large language models) have come under fire for their excessive energy consumption and consequent ecological footprint, with some researchers estimating that LLMs consume hundreds of liters of fresh water and produce annual emissions equivalent to those of a small country.
See also: Generative AI: It alone will not improve the customer experience
With global warming exceeding 1.5 degrees Celsius for the first time in a full year, global stakeholders are debating where the ultimate responsibility for preventing the climate crisis from worsening should lie. Climate change remains a critical issue for both consumers and companies in these global efforts to reduce CO2 emissions, which does little to stop the common perception of a company using AI consumption tools without controlling the carbon it leaves behind. More importantly, the rampant, uncontrolled use of AI could have devastating consequences for the environment – research from MIT suggests that education alone has the potential to significantly reverse global progress in the fight against climate change.
Despite the obvious ecological indifference of recent legislative measures such as the EU law on artificial intelligence and President Biden's directive, which focus mainly on other aspects of AI responsibility, some major AI players have begun actively self‑regulating and working toward sustainable AI use. Here are ways in which AI leaders approach AI with ecological consciousness while maintaining the deep business value of the technology.
See also: BBC: Creating an Internal Artificial Intelligence Model?
Specifically designed artificial intelligence
Many disadvantages of generative artificial intelligence and LLMs stem from the massive data reserves that must be navigated to produce value. This not only increases risks regarding ethics, accuracy, and privacy, but also significantly raises the amount of energy required to use the tools.
Instead of highly general AI tools, businesses have begun to move towards more narrowly specialized AI tools, specialized for specific tasks and goals. For example, ABBYY has taken this approach by training machine learning and natural language processing models to read and understand documents flowing through enterprise systems like a human. With AI skills pre-trained to process highly specific document types with 95% accuracy, organizations can save trees by eliminating paper usage, while also reducing the amount of carbon emitted through tedious document management processes.
See also: Melting ice will change time as we know it
Empowerment of developers
AI companies do not need to shoulder the burden of sustainable AI themselves – some are actively trying to throw the proverbial ball onto the developers' field.
OpenAI , the AI pioneer behind the popular ChatGPT , recently announced that developers can create their own “GPT” platforms for specialized purposes. This allows developers and organizations to narrow down their use of highly customizable AI, removing unnecessary features and data that exacerbate ecological damage. For example, developers could design GPTs for purposes limited to creative writing tips, cooking information, tech support, or any other specialized purpose.
Considering the increased risks for inaccuracy and privacy violations associated with highly general AI models, developers are likely interested in leveraging these more limited, more specialized GPT platforms not only for their ecological responsibility, but also for improved business outcomes.
See also: The end of space travel as you knew it – The Dream Chaser is here!
Sustainable business practices
Companies must also look beyond the technology itself and search within their organization for more ways to sustainably exploit artificial intelligence. For example, Microsoft revealed that the hardware supporting artificial intelligence operates exclusively on clean energy, freeing them from the creation of the so‑called “operational emissions”.
Additionally, companies can use artificial intelligence as a tool to explore other aspects of their business where sustainability could be prioritized. Forrester highlights the measurement, reporting, and data‑visualization capabilities of artificial intelligence to indicate that it could drive a climate revolution on its own.
Although objectively significant, emissions are not the only metric used to cover ecological impacts – studies have shown that a combination of robotics and artificial intelligence has reduced the use of herbicides in some contexts by 90%. As companies continue to engage with the usefulness and consequences of artificial intelligence, they must explore the full range of its capabilities to strengthen and contribute to sustainability.
See also: Gaia reveals ancient streams of stars in our Milky Way
Businesses take responsibility
So far, legislation on artificial intelligence has largely failed to control the ecological impacts of artificial intelligence, focusing instead on privacy and other ethical areas. While these areas are also critical for the responsible use of artificial intelligence, businesses must remain accountable for the way they leverage artificial intelligence to create business value.
The year 2023 may have been a year of excessive hype, noise, expectations and misunderstandings around artificial intelligence, but the maturity that businesses have gained over the past year gave them the means they need to make informed and responsible decisions regarding the use of artificial intelligence. Nevertheless, it is wise to examine, challenge and hold large organizations accountable for the carbon they leave behind and other environmental impacts – those who prefer their ecological responsibility should have nothing to hide.
Source: techradar
