AI received over $250 billion in global investment in 2024, with that number only increasing over time.
This money is used to enhance the intelligence and realism of AI, improving images, videos, speed, learning capabilities, and increasing its overall usage.
Most recently, the app Sora 2 was released to beat every other AI platform in video and image generation, having incredibly realistic graphics and physics.
All of this sounds great; however, there are multiple dangers hidden behind AI’s broad capabilities.
As reported by CBS News, Hany Farid, professor of Electrical Engineering and Computer Sciences at UC Berkeley, said, “Anybody with a keyboard and internet connection will be able to create a video of anybody saying or doing anything they want.”
And that is where many problems with AI lie: there is little to no restriction on how much people use it or how it is being used.
The Times stated that “Users can simply type in text to the generators to create a fake and explicit image” in many AIs, revealing the possibilities of misuse.
According to Wired, when an AI image generator’s database was exposed, it was discovered that there were “tens of thousands of explicit AI-generated images” allowed to be made, due to a lack of restrictions.
However, since 2024, companies and governments have imposed new restrictions to stop the generation of explicit or illegal imagery, such as Athropic’s idea to make a system called “constitution classifiers,” which can “monitor both inputs and outputs for harmful content,” remarked Financial Times.
However, restriction isn’t only for wrongful use; it’s also for preventing overuse.
According to The Guardian, “Unrestricted AI use could add [one billion] tons of planet-heating emissions in the US over the next decade.”
One AI app alone uses an excessive amount of electricity, so as AI becomes increasingly widespread, electricity consumption will increase further.
On top of that, AI systems need water to cool themselves.
As stated by Forbes, “AI server cooling consumes significant water, with data centers using cooling towers and air mechanisms to dissipate heat, causing up to 9 liters of water to evaporate per [kilowatt hour] of energy used.”
With 3.2 liters of water being consumed per person every day on average, the amount of water that AI uses will stack up fast.
Gianluca Guidi, visiting scholar at the Harvard TH Chan School of Public Health, remarked, “We talk a lot about what AI can do for us, but not nearly enough about what it’s doing to the planet.”
However, as reported by PWC, “New analysis suggests that in the decade ahead, AI could boost energy efficiency enough to save as much energy as the technology uses.”
Something else that raises concern about AI is the common notion that it could someday take over humanity.
Now, although it’s “smart,” humans still hold the power over AI.
As Forbes stated, “While AI can process vast amounts of data and identify patterns, it lacks the nuanced understanding that comes naturally to humans.”
Since humans have control over AI, which has only a non-sentient thought process, there is no chance of it rebelling on its own.
However, since humans have control over AI, who’s to say that it won’t fall into the wrong hands and go haywire?
One extreme example of AI being misused is Microsoft’s AI chatbot Tay. “It took just two tweets for an internet troll going by the name of Ryan Poole to get Tay to become antisemitic,” The Guardian reported.
The reason this happened is because Tay was made to learn from the people on X, formerly Twitter, and other platforms, causing her to be easily misinformed.
As stated by IEEE Spectrum, “Tay’s [built-in] capacity to learn meant that she internalized some of the language she was taught by the trolls, and repeated it unprompted.”
However, as AI gets better, testing is put into place, such as the “red teaming” technique, where “Teams of human testers write prompts aimed at triggering unsafe or toxic text from the model being tested,” as put by MIT News.
All things considered, the current state of AI may not be ideal, and it can be manipulated in many ways; however, it’s worth noting that as AI progresses, it will be optimized to have better restrictions, learning systems, energy usage, and be a better tool to use in everyday life to help people, not replace them.
But that will only happen if humans can learn to work with and understand AI.