The AI chatbot war is in full swing

Google parent company Alphabet might as well change its name to Alexander, because it had a terrible, horrible, no good, very bad day yesterday. Its stock dropped nearly 8% – erasing more than $100 billion in market value – following the company's AI- and- search-focused event held in the morning.

And to add insult to injury: someone noticed Google's AI made a rather large error in a Twitter post… that’s now been viewed more than 1.6 million times.

Some highlights from the event:

• The company mainly rehashed Monday’s announcement of Bard, its AI-assisted search engine product, showcasing how it could be used to display the pros and cons of buying an electric car, as well as plan a trip in Northern California.
• The event also showed AI improvements to a number of other Google products, including Maps and Google Lens, which lets people search for images from their phone’s camera.

🪐🤦 Now, about the error: Right before the event began, Reuters discovered Bard making a factual error during a demo Google itself posted to Twitter two days prior. When asked how to explain the James Webb Space Telescope to a nine-year-old, Bard incorrectly said the JWST “took the very first pictures of a planet outside of our own solar system.” Except… it didn’t; those pictures were taken in 2004 – 17 years before the JWST ever existed.

• This also showcases a bigger AI chatbot issue: the “hallucination problem.” Because the AI’s language model is trained to basically be a super powerful autocomplete, it doesn’t check to make sure that its answers are factually correct – only that they’re probabilistically so.

🖼️ Big picture: AI’s answers may not always be accurate, but the following statement is: the chatbot wars are heating up. Microsoft, which has invested billions in ChatGPT’s parent company, OpenAI, officially unveiled its own AI-enhanced Bing search engine with ChatGPT integration on Tuesday.