AI As a Growing Political Concern

Artificial intelligence is no longer just a topic of national debate; it’s becoming increasingly relevant at the local level, too. Across the country, communities are now striving to attract the substantial data centers essential for fueling the AI surge. However, this rush raises alarms among environmental activists about water usage, while ratepayers express anxiety over potential spikes in their electricity costs. Additionally, there are worries among workers that advancing technology may threaten their jobs.

The pushback against AI is not limited to a single political faction, and many of the issues raised are indeed worth serious consideration. However, conversations often start and end with alarming statistics devoid of the context necessary for real understanding.

As AI progresses, the United States faces some hard decisions ahead, and it’s vital to base these on facts rather than numbers that merely stoke fear.

There are several pertinent questions regarding AI, such as issues of privacy, surveillance, censorship, individual autonomy, and the concentration of power. Nevertheless, the three main concerns often voiced about AI and data centers tend to be exaggerated.

Claim 1: Data Centers Use Excessive Water

Reports about data center water consumption tend to begin with eye-popping figures. Some facilities are reportedly designed to use millions of gallons daily, while others might consume billions annually. In areas grappling with water shortages, such figures can indeed be alarming.

Yet, context matters greatly. An analysis suggested that, based on 2021 statistics, all data centers in the U.S. together are estimated to account for merely 0.3% to 0.4% of the nation’s daily water withdrawals.

Furthermore, cooling methods vary significantly. Some data centers primarily use evaporative cooling, while others utilize closed-loop and air-cooled systems that require much less water. Factors like the facility’s location, water source, and cooling design can all impact water usage greatly.

For perspective, data centers are not unique in their water consumption. Golf courses across the U.S. use about 30 times more water annually, and California’s almond farms are estimated to utilize over 100 times more.

Of course, poorly placed data centers can still create pressure on local water supplies. So merely stating “millions of gallons” isn’t a definitive argument; one should also consider whether a given facility significantly burdens an area’s water system.

Claim 2: Data Centers Consume All Available Power

The energy consumption of data centers is undeniably significant. Projections indicate that, by 2025, U.S. data centers will be set to use around 200 terawatt-hours of electricity, which would be about 4.7% of the country’s total electricity use.

However, critics jump to conclusions that increasing demand for these centers would inevitably raise household electricity costs. But the reality isn’t that straightforward. Recent research shows that despite growth in data centers, their effect on retail rates has been limited, with broader inflation and energy policies being more influential in consumer price increases.

This doesn’t imply that new data centers won’t impact the power grid. Unsuitable rate structures might shift costs onto consumers, and poorly planned facilities could put a strain on regional power generation. Yet, the answer to rising demand isn’t simply to suppress it; instead, it’s about enhancing and diversifying energy sources.

The AI boom may actually be creating that very incentive to improve infrastructure. For instance, Amazon’s recent acquisition of an 8,000-acre site in Texas for a data center could facilitate up to 7.65 gigawatts of on-site natural gas energy generation. Likewise, Microsoft is looking to harness power from the Three Mile Island nuclear facility.

While it’s been historically tough to construct reliable energy systems in the U.S., the current demand for AI might be the catalyst needed to instigate a shift in that trend.

Claim 3: AI Will Replace All Jobs

Among the claims surrounding AI, the prospect of job loss is particularly striking. It’s true that AI has the potential to eliminate specific roles, particularly entry-level positions. Long-term predictions are tricky, mostly because it’s unclear how efficiently these technologies will operate. Still, history gives us pause against trusting these dire forecasts uncritically.

Back in 2016, AI pioneer Jeffrey Hinton suggested that people should stop training to be radiologists since machines were about to exceed human capabilities in reading medical images. Fast forward a decade, and radiology isn’t just surviving; the number of U.S. radiologists has increased, alongside rising salaries and workloads due to advancements in medical imaging.

This could exemplify what’s known as the “Jevons paradox.” When technology enhances the efficiency of resources and services, overall demand often increases. Recent studies have applied this theory concerning AI and healthcare professionals.

The same logic can apply elsewhere. If AI dramatically boosts productivity for programmers, it might seem logical to think the economy would need far fewer programmers. But the economy often adapts; companies that couldn’t afford custom software may suddenly find it viable, unleashing new products and services along the way.

None of this guarantees smooth transitions for all displaced workers, and not every profession will thrive. Technological advancements will inevitably favor some over others. However, the essential point remains: productivity spikes don’t automatically equate to equivalent job losses.

Broader Concerns

This doesn’t imply that AI is devoid of risks. In fact, the most pertinent concerns might not align with those making headlines.

As AI continues to weave itself into the fabric of society, these systems could influence the information people access, the decisions made by institutions, and the ways individuals interact with governments and businesses. Collecting vast amounts of data may lead to heightened surveillance, and biases could seep into algorithms. An increased reliance on AI might erode personal autonomy, centralizing decision-making in the hands of a few corporations and governments.

These issues deserve more attention than sensational claims concerning water usage, power, and job security.

As AI progresses, America will face some tough choices. The smart approach is to ground discussions in facts rather than figures designed to incite alarm. The ultimate goal should focus on safeguarding freedoms, privacy, and individual rights while nurturing innovation. Overselling simpler issues can inadvertently complicate the more challenging ones that require addressing.