Google's plan to respond to the ChatGPT threat is not convincing investors, triggering strong selling pressure on the shares of parent company Alphabet this Wednesday and taking 100 billion dollars off the company's market value, according to Reuters.

The group's shares closed the session down around 7.5%, but were down 9% at one point, after it was made public that a Google promotional ad for the new Bard chatbot shows the program giving a factually wrong answer to a question.

This Wednesday, Google presented new features based on artificial intelligence, at a time when competitor Microsoft has been getting closer to ChatGPT, a model developed by OpenAI capable of answering human questions in natural language and which has reached 100 million users in two months of existence.

Google's alternative is called Bard, but an announcement shared by the company on Twitter contains a factual error. The news was reported by Reuters and cited by ECO, and represents a humiliation for the company, while also shedding light on the risks inherent in this type of technology.

According to the agency, Google's promotional ad (in the tweet below) shows Bard answering a question incorrectly. To the question "What new discoveries from the James Webb Space Telescope can I tell my nine-year-old?", the program returns a list of options, including that the telescope in question was used to take the first photographs of a planet outside the solar system. But Reuters confirmed through information from NASA that the first images of these exoplanets were taken by the European Southern Observatory's Very Large Relescope in 2004.

Confronted with this situation, an official Google source told Reuters that this case "highlights the importance of rigorous testing processes", which were initiated "this week". "We're combining external feedback with our own internal testing to ensure that Bard's responses meet a high standard of quality, security and grounding in real-world information."

The international press has reported that the rapid growth in popularity of ChatGPT has set alarm bells ringing at Google. ChatGPT is free and responds to user requests, but OpenAI points out that the program has no information on events that occurred after 2021. In addition, the ChatGPT model also makes factual errors in its responses, something that artificial intelligence experts call "hallucinations".