The Gen-AI bubble and its bursting point

There is a sense of wonder when it comes to the invention of generative artificial intelligence (Gen-AI). Being able to interact with machines as if we were interacting with another human was something we could only see in science-fiction films when we were children. Yet, over the past three years, we have seen immense advances in Gen-AI – to the point where it is increasingly difficult to distinguish between interactions and content that are real and those generated by AI.

Indeed, Gen-AI is a remarkable product of technological advancement. AI, and therefore Gen-AI, is a tool that can improve efficiency, automate interactions with customers and prospects, and potentially reduce labour costs. However, over the past three years, we have also been reminded that products such as plastic and asbestos were once considered "wonder products" until we learned about their harmful effects.

Artificial intelligence itself is not a new concept. We have long seen it in large-scale machinery, such as the robotic arms that assemble cars, as well as in small devices such as the predictive-text functions on our smartphones. There have also been remarkable breakthroughs in life-saving technologies, including in surgery. AI has long existed as a tool for humans to use.

But with the introduction of Gen-AI, we have seen a disruption in technology on a scale that was previously difficult to imagine: machines can now generate responses that appear to "think" for themselves.

In simple terms, Gen-AI processes vast amounts of data and generates a response based on the prompt provided by the user. With so much information available online, Gen-AI systems can be trained to produce increasingly sophisticated responses to user commands.

But as time went on, publicly available data became more limited, and technology companies and other corporations increasingly began looking towards personal data and individuals' intellectual property for training purposes. As a result, many social media, art and broadcasting platforms have introduced terms of service that allow AI systems to learn from and train on users' content.

Meta, the owner of social media platforms such as Facebook, Instagram and WhatsApp, was reported to have used users' personal information to train AI for "marketing purposes" as early as November 2024, according to The Guardian and Al Jazeera. Companies that host platforms for artists, musicians and other creators – such as Spotify, DeviantArt and YouTube – have also introduced policies allowing creators' content, old and new, to be used for AI training.

For creators who have few alternative platforms with comparable visibility or reach, opting out can be difficult. In practice, this can leave them with little choice but to accept policies that allow their work to be used in AI development.

Additionally, several studies have raised concerns about Gen-AI's potential effects on cognitive skills and self-reliance in everyday tasks. A recent study discussed by the Harvard Gazette pointed out that while Gen-AI can open up new avenues for self-education, some users may instead come to rely on it as a crutch rather than use it as a tool. Over the past few years, over-reliance on AI and plagiarism in academia have also become growing concerns.

Moreover, there have been several cases in which Gen-AI has generated fabricated "facts" when users rely on it for fact-checking. This can undermine the quality of the information ecosystem on which society depends. Gen-AI systems cannot always reliably distinguish between different types or levels of source credibility, meaning they may draw on opinion blogs or other less authoritative sources when responding to serious questions.

Gen-AI also has significant environmental implications because of the data centres required to run these systems. Data centres can require substantial quantities of freshwater to maintain suitable temperatures through cooling systems, while also consuming large amounts of electricity to operate their servers.

A recent report by CNN discussed AI data centres creating "heat islands", with areas within a radius of around 10km reportedly experiencing temperature increases of between 2°C and 8°C in extreme cases. Another report by the BBC highlighted concerns over the large quantities of water used to cool data-centre systems and the potential impact on freshwater resources.

If such impacts are being reported in parts of Europe and the United States, it is worth asking what the consequences may be in parts of Asia and Africa, where the environmental impacts of data centres may receive less scrutiny.

The labour market has also been affected amid the AI boom. Over the past 18 months, there has been a significant wave of layoffs as corporations explored replacing human labour with AI.

Interestingly, some corporations that laid off thousands of employees in an effort to replace human labour with AI are now reconsidering those decisions. Recent reports involving companies such as Microsoft and Uber suggest that some businesses have found that AI adoption alone can be more costly or less effective than initially expected.

A recent report by Forbes said 29% of employees laid off by corporations and businesses in an effort to replace them with AI over the past 18 months had subsequently been rehired. The report also said 55% of executives regretted trying to replace employees with AI rather than restructuring tasks to integrate human labour with AI. The rehiring trend could increase further.

AI is undoubtedly a powerful tool that could help advance humanity. Yet it remains unclear what advantages Gen-AI, in its current form, is ultimately supposed to provide beyond the efficiencies it already offers.

Some may compare Gen-AI with the invention of the car. I would personally compare it to the invention of plastic – a remarkable product that needs to be handled carefully rather than used unsustainably.

Gen-AI has the potential to contribute to further advances in AI and, ultimately, humanity. But we are yet to see that potential fully realised – at least for now.

Araf Momen Aka is an executive of Education Development Program at Radiant Business Consortium Ltd.

Disclaimer: The views and opinions expressed in this article are those of the author and do not necessarily reflect the opinions and views of The Business Standard.