Artificial Intelligence and supply chains

Artificial Intelligence and Supply chains

Modern Artificial Intelligence, such as ChatGPT and Copilot, are great tools, however they create stresses and strains on other industries and sectors in some unexpected way.

When training Large Language Models, the technology which enables tools like ChatGPT, huge amounts of computing resources are required. As we touched on earlier, this consumes a massive amount of energy and the training requires vast data centers filled with advanced hardware. What are the consequences of this other than the obvious environmental impact? 

Artificial intelligence’s hunger for energy

As you read in the previous step, the training of artificial intelligence systems requires an immense amount of energy. There are many discussions to be had about whether this is responsible, or if sufficient value is derived from the training. However, a more serious issue arises. As many of the world’s nations have made commitments to reduce carbon emissions, and move towards renewable and green sources of energy, some energy producers are doubling down on the use of coal, oil and gas energy to meet the requirements of data centers that are training AI models.

You can read more about how the requirements of AI energy demands are impacting energy generation.

Consider the following, and add your thoughts to the Comments: 

• Is it responsible to turn our backs on carbon emissions targets to train AI models?

Another recent development is the potential reopening of three mile island nuclear plant reopening after a number of years. The Three Mile Island incident was a partial nuclear meltdown that occurred on March 28, 1979, at the Three Mile Island Nuclear Generating Station in Pennsylvania, USA. It is considered the most serious accident in the history of the American commercial nuclear power industry, though it did not result in deaths or injuries to plant workers or the public.

You can read more about the reopening of Three mile island nuclear power plant.

Consider the following, and add your thoughts to the Comments: 

• Is it responsible to use contentious nuclear power to train AI models?

Artificial intelligence’s impact on supply chains

Data centers that are used to train AI models require the latest developments in technology. This includes the latest CPUs, GPUs and switching infrastructure. As companies battle to train larger and larger models to gain a competitive advantage, they are commissioning huge data centers, consuming more equipment than manufacturers can build. The consequence is that other industries are starved of resources, and as you would expect, when demand outstrips supply, the price of the product sky rockets.