Alarm bells have been ringing about the potential risks posed by the rapidly expanding artificial intelligence (AI) technology.
In a warning of what can happen when businesses over-rely on AI, on Thursday, a global outage of ChatGPT occurred.
The risks of AI are manifold and range from short-term concerns like the loss of millions of jobs, to longer-term existential risks to humanity — concerns which have increasingly climbed up the list of worries of politicians and scientists alike.
Despite these concerns, investors continue to support AI technology. After NVIDIA Corporation’s latest earnings call, the company’s stocks soared by 26%, thanks to its reputation as an “unquestionable AI leader.”
The Hazards of Excessive AI in the Workplace
According to a research paper released by OpenAI’s research arm, up to 80% of the US workforce could see up to 10% of their work tasks affected by the introduction of GPTs — with some workers potentially seeing half of their roles impacted. These impacts are expected to hit various levels of the wage ladder and cut across different industries, with higher-income jobs potentially facing greater exposure.
But there’s an added hazard which many disaster predictions have failed to consider: the risk of over-relying on AI in the business world.
Thursday’s outage of ChatGPT, which lasted just over an hour from around 7.23 AM Eastern Time to 8.27 AM, per Gizmodo, serves as a reminder of this additional danger.
ChatGPT is typically accessed via OpenAI’s server. While it is possible to operate the program via a private server, OpenAI’s business model relies on subscriptions, so most companies — even those with adequate infrastructure — benefit from not having to host the AI on their own servers.
As more companies rely on ChatGPT to complement the development of chatbots and digital assistants, as well as for data classification, text analysis, and coding assistance, the potential for a ChatGPT outage to result in an operational outage across firms becomes a more significant concern.
ChatGPT’s text-generation capabilities have also become crucial for many businesses in brainstorming ideas, writing copy, creating social media content, writing presentations, emails, and more.
While these tasks are still ultimately carried out by humans and can be replaced by them in the event of AI failure, the implications of an over-reliance on AI in the long run can lead to productivity drops across the board if similar technologies become ubiquitous.
Moreover, an over-reliance on automated text generation could land firms in legal trouble, from the output of false data and information to the perils of gender or race-based discrimination. Uploading tasks to ChatGPT-like software can lead to companies making unwanted errors that snowball into legal hazards.