This week has been decisive in the so-called Artificial Intelligence War. Ever since OpenAI launched ChatGPT and Microsoft acquired its formula with a $10 billion investment, the world has been waiting for Google to make a move, given the threat it posed to the search engine. And finally, Google has announced Google Bard, its new AI that will be a turning point for the search engine.
But not everything went as smoothly as expected. Almost at the same time that Google launched Bard, Microsoft trumpeted Prometheus Model, the integration of ChatGPT into both Bing and Microsoft Edge. The announcement has come with the consolidation that it will use the GPT-4 language system, the latest update of the technology. And it has done so alongside the promise that “it’s a new day for search engines”.
Both Bard and Prometheus, a priori, do similar things. In the end, they are AIs that will serve to provide rapid responses to users and will use the information gathered by the Internet in real time to be as effective as possible. However, there is a clear winner in this first battle: Microsoft has clearly beaten Google.
You only have to look at the graph of the change in its shares after the announcement of Bard and Prometheus to realize the “brutal beating” that Bill Gates’ company has given to Sundar Pichai’s company in its presentation.
While Microsoft has grown exponentially and established itself in much higher numbers, Google has plummeted to much lower figures than usual. Considering that they were starting from a virtually level playing field, it is clear that the two announcements have had very mixed reviews.
What is the problem with Google Bard?
The main reason why Google Bard has not worked as well as expected in its presentation is based on a problem of choice of examples about the tool. In a Google blog post advertising Bard and talking about the benefits of this new language model, several examples of answers to simple questions on various topics were published.
One of the questions that could be read was “what new discoveries from the James Webb Space Telescope can I tell my 9-year-old son?”. The question, of course, includes some interpretive subjectivity that a normal searcher would not be able to manage. One of the answers generated to such a question about the NASA telescope reads “JWST took the first images of a planet outside our own solar system.” And… no, it is not true. The first image taken in the history of mankind of an exoplanet was captured by the Very Large Telescope in Chile, in the year 2044.
This failure has caused a decline from $106.77 to $98.98 for Google’s parent company, Alphabet. This translates, to give you an idea of the size of the problem, into a loss of $120 billion. And if you add in the fact that Microsoft was succeeding in the meantime, you can imagine the drama.
This error has led to mistrust of Bard. If not even Google employees are able to provide good examples of how it works, what will the user experience be like? Okay, it is clear that it is an AI that still needs to be perfected and probably won’t fail so much in the future. But ChatGPT has been winning over users since its inception, so Microsoft had it made.
Still, this is only the first battle of many to come in the AI War. So we’ll be watching to see how it all evolves.
