When generative AI tools first became available to the general public, there was an unparalleled blend of amazement and apprehension. These tools could craft stunning images and articulate words from mere text prompts, sparking Silicon Valley’s interest. After the underwhelming performance of cryptocurrencies and the metaverse, the tech industry saw generative AI as a potential goldmine. However, alongside this enthusiasm were fears of job losses, difficulty distinguishing between AI and human-generated content, and, for the more dramatic, concerns about humanity’s very survival.
Fast forward a few months, and the charm seems to be waning. Governments are moving to regulate generative AI, intellectual property disputes are on the rise, and questions about user privacy have emerged. Moreover, there are increasing doubts about the accuracy and reliability of AI chatbots. The initial craze appears to be settling, and some reports suggest that the public’s interest is diminishing.
OpenAI’s ChatGPT was a forerunner in the generative AI buzz, largely due to its early public release and a significant partnership with Microsoft. This collaboration saw Microsoft infuse Bing, its search engine, with a chatbot powered by OpenAI’s technology, promising users a richer, more integrated response to their searches.
However, recent data indicates that this ambitious move hasn’t notably impacted Google’s dominance. Even with the hype, Bing’s market share remains nearly unchanged. Microsoft remains hopeful, planning to make Bing Chat accessible on more browsers, which could potentially boost its popularity.
OpenAI’s ChatGPT also seems to be experiencing a slight downturn. For the first time since its launch, there’s been a noticeable decrease in website traffic and app downloads.
In response, Google introduced Bard, its chatbot, but has been careful about its integration, keeping it separate from its primary search platform. Google’s caution is understandable given the several instances where chatbots have shown glaring flaws, from producing erroneous information to exhibiting biases.
A concerning revelation is that OpenAI’s latest model, GPT-4, seems to be showing reduced accuracy over time. Some attempts to use AI for journalism have also resulted in glaring mistakes.
In a collaborative effort, eight major AI companies, including OpenAI, Google, and Meta, recently participated in DEF CON, a large hacker convention, to test their models for accuracy and safety. The event, supported by the Biden administration, aimed to identify vulnerabilities in these AI systems.
Despite concerns, generative AI’s potential is undeniable. Its initial allure drew immense attention from both the tech industry and the public. The challenge now is to refine and enhance this technology, making it more than just a fleeting gimmick.