Should Google care about ChatGPT?

In late November, OpenAI, a nonprofit AI research organization, was introduced to the public gpt chat, a program designed to talk to users through chat. This type of tool, known as a “chatbot”, has been around for quite some time, but ChatGPT has become famous for its great ability to generate textual content of all kinds, producing highly reliable and amazing answers.

From the early days of experimenting with ChatGPT, many users have noticed that the expressive capabilities of a chatbot go far beyond generating more or less realistic answers: thanks to the variety of contents on which the model is trained, it is often able to provide useful answers to questions of various kinds. newly item Posted by The Atlantic has come to warn of the consequences mass use of similar services could have in teaching, given that ChatGPT today seems capable of generating a short article in a few seconds good enough to get by – with some modification by the user.

– Also read: A chatbot like no other has arrived

The fact that it is enough to type a short command and press a button to receive an answer on any (or almost) topic has led some experts to assume that artificial intelligence like ChatGPT could pose a significant threat to Google and other search engines, which at the end of the 90s organize and distribute content on the Web. Indeed, in the future, asking a chatbot for advice and information may become faster and more practical than entering a series of search terms into the Google bar, hoping to get the desired answer from among the many results.

The new capabilities offered by artificial intelligence are reaching a sensitive stage for the search engine, which has been around for several years Criticize The increasing importance given to ads in displaying search results. With the consequent deterioration of the user experience. In addition to the increase in advertising, what greatly worsens the quality of searches on Google is the standardization of a good part of the content on the web, often produced according to the stylistic rules of SEO (search engine optimization), in the hope of being selected and rewarded by the algorithm . As a result, searches on Google are much less relevant and less accurate than they used to be, and it’s even more difficult for users to find what they’re looking for. Also for this reason, there are those who believe that products like ChatGPT could point to an important alternative to searching the net.

– Also read: Searches on Google are not what they used to be

Just last week he was an analyst at US investment bank Morgan Stanley size The threat AI systems seem to pose to Google. However, not all industry experts are so optimistic: according to The New York Times, within Alphabet, the group that includes Google (but also YouTube and other companies), some executives are very concerned and may announce “Code Red”, an euphemism for sounding an alarm internally to give top priority to a potential issue or threat. According to the newspaper, in fact, “some fear the approach of a moment that major groups in Silicon Valley feared for the company: the arrival of a massive technological change capable of disrupting business.”

However, it would be wrong to say that Google was surprised by the emergence of OpenAI or other services such as MidJourney AI and Stable Diffusion, which are widely used in creating images. The company was an early investor in the industry and has already built a chatbot that it believes can compete with ChatGPT. Google also has a privileged relationship with the GPT language model itself, from which it developed part of the technology (more precisely those related to the “T” from the GPT acronym, or the “Transformer” neural network, Foot in 2017).

The chatbot developed by Google has been talked about in recent months in connection with controversy that has arisen precisely from the great communication and articulation skills demonstrated by the technology. Last June, programmer Blake Lemoine, who worked in the artificial intelligence department at Google, told the press that he had evidence that LaMDA, the discursive artificial intelligence developed by the company, was so advanced that it to be Indeed “conscious”.

Lemoine’s remarks caused a sensation but were quickly refuted by analysis of the conversation between the programmer and the chatbot, by which it became clear that LaMDA responded to user commands with only the impression of providing answers that appeared to come from a talented individual. with self awareness. Lemoine was fired from Google and the case was interpreted as evidence of the great capabilities of LaMDA. Despite this, the company is still reluctant to open its use to the public, as OpenAI did with ChatGPT, making it only available to a few researchers and scientists.

It is precisely this degree of caution that separates Google from OpenAI, guaranteeing the latter a huge advantage, at least in terms of media. Similarly, Google has always argued that the future of search is “conversational”, that is, the simple searches we know today will gradually be replaced by a continuous dialogue between the user and the service. The transition to this new type of search is being hampered by two major hurdles: the first has to do with Google’s size and relevance in the digital world; The second relates to the advertising sector, which is central to the group’s finances.

– Also read: How artificial intelligence designs the world

Yet, the texts ChatGPT produces, while very convincing, often contain errors, and the speed with which the service can produce texts of any kind, from poems to science hoaxes, has many commentators concerned. According to them, the popularity of these artificial intelligence systems will have serious consequences in the spread of disinformation and fake news, to the point that someone comparison to the “uncontrolled release of the virus into the environment.” OpenAI co-founder and CEO Sam Altman himself has it Known The service is “incredibly limited but good enough in certain things to create a false impression of cool”. Two years ago, a few months after the introduction of GPT-3, he had Known You exaggerate Expectations related to technology and highlighting its flaws and the many mistakes made.

The ultimate goal of ChatGPT, at least for now, is to create meaningful written answers, not to be an oracle capable of giving the right answer to any question. This is a common limitation of any such product, including LaMDA, and it’s what makes widespread implementation problematic by Google, which can’t afford to offer an unreliable, if surprising, service.

Even if Google could solve such a complex problem, the advertising question would still be open. The search engine’s business model involves showing ads near search results, and the company hasn’t yet figured out how to do the same in a conversational model, where ads have to interrupt between the user and the device. According to journalist Alex Kantrowitz“So Google has little incentive to move beyond traditional search, at least not in a paradigm-shifting way, until it figures out how to make the economy work. In the meantime, it will continue to focus on Google Assistant,” its voice assistant.

Speaking of money, ChatGPT today is a very expensive shareware product. On OpenAI’s user mark of 1 million users, Altman said the cost of maintaining and operating the system “brings tears to your eyes.” The company has received investment from several companies and investors, including Elon Musk and Reed Hastings of Netflix, but is particularly dependent on Azure, the division of Microsoft that handles Cloud computing and digital infrastructure.

The use of a service like ChatGPT by a large, albeit limited, audience has a significant cost, which according to some estimates will be Three million dollars Per day. It is also why, according to some observers, the economic side will soon end this season of free trials, forcing OpenAI and other companies in the sector to set a price for their services.