AI bots talking to bots may mark the start of the internet’s mid-life crisis

livemint.com:

Subscribe to enjoy similar stories. Till recently, the digital frontier looked boundless, a perfect virtual world where human interaction and the exchange of ideas were encouraged and democratized. With the rise of AI tools, the internet is witnessing several disconcerting trends.

An Amazon Web Services (AWS) study reveals that over 57% of sentences on the web have been translated into two or more languages with machine learning. Users in Africa and South Asia encounter content that has been poorly translated or distorted into a misleading click-bait. The motive? Advertising dollars.

The underlying problem is a combination of tech limits and global inequalities. In low-resource regions, where there is insufficient linguistic data available to train AI models, translations tend to be wildly inaccurate. As a result, reality is distorted for profit, with users left without recourse to redressal.

One disorienting theory in circulation is the idea of the ‘dead internet.’ Part conspiracy theory, part science fiction, this posits that most online interactions are no longer between human beings, but are dominated by AI bots. These bots are responsible for everything from fake social media profiles to orchestrated political narratives. The theory also suggests that governments have leveraged this technology to create a vast mind-control mechanism, filtering what we see and shaping the public discourse to suit their needs.

While this may be exaggerated, it does suggest a problem. The internet has become a battleground between tech companies and those looking to exploit their algorithms. In 2013, YouTube employees found that this platform was being gamed by users artificially inflating likes and views.