As AI services, software, and elements become more and more commonplace, it is worth examining the side effects that are coming to light. The adverse effects of AI usage are extensive. From climate impacts to societal shifts to personal and professional risk, we have a lot to consider already, and we are only at the beginning of this new technological age.

Possibly the most talked-about side effect of artificial intelligence is the cost to the environment. AI models require loads of computing power, and all that power comes at a price. The amount of electricity needed to run major computational facilities and the water necessary to keep those facilities cool is concerning. If you’ve ever been in a server room, you know that high-performance machines emit a lot of heat. The cheapest way to keep a massive facility working at peak performance is to use a local water source to cool your space. While the financial output may be low, the cost to the planet is high.

NPR recently reported on a consequence of “zero click searches,” a term coined to refer to the instances when someone Google searches a question and then gets their information from the AI overview supplied at the top of the page, thereby negating their need to read more or click on any of the links below. Hence: “zero click search.” Studies have shown that the rise of these AI overviews is directly connected to a drop in readership on journalistic websites of all stripes. No matter the content – from the Wall Street Journal to Rolling Stone to Wired – organizations sharing the written work of human journalists are seeing a sharp decline in clicks. It will not take long for that decline to show up in their advertising revenue, and eventually, their staffing budget. The societal ramifications of losing journalists are huge.

Lastly, a May court order decreed that OpenAI must keep logs of all ChatGPT’s conversations. This is something that could affect all of the individuals using ChatGPT as a sounding board. All stored records, by virtue of their continued existence, are subpoena-able and can be used in court cases. This is something all users should consider when using the program. Whether you are asking ChatGPT to spruce up a work email, diagnose a personal malady, or provide tips on disposing of evidence of a crime, all of that can be accessed long after you have deleted it on your end. This could change the way criminal cases are prosecuted for years to come, much like the rise in GPS location data and text messaging communication. It could affect investigations into wrongdoing at work, civil cases, or insurance claims. Anyone who shares sensitive data with ChatGPT needs to be prepared for that information to become public.

Machine learning and generalized AI models are exciting and fast-changing technologies. Here at SpitzerTech, we love cutting-edge tech, and we are always curious to see what’s next. In the case of AI, we are pairing that curiosity with a dose of caution, and we think you should as well. If you want to talk through the risk/benefit ratio of any AI products you are considering incorporating into your business, give us a ring. We will help you look at the WHOLE picture, and identify risks and liabilities for you and your organization.

X

This field is for validation purposes and should be left unchanged.
Name(Required)
Please let us know how we can help.