Label

Recent

Bing's AI is Off to a Very Weird Start

Advertisemen 300x250

With the sudden and fierce rise of AI technology in recent months, various companies like Microsoft and Google are looking to capitalize on its potential with their takes on automated content creation. However, Microsoft's AI tool to support its long-running search engine Bing is having a rough and very odd start in embracing artificial intelligence.

Microsoft has announced an update to the Bing AI chat, and it will now output shorter responses in light of the oddball interactions users have had with the tool. The developers cited that the overly long posts that the AI tool would put out ended up resulting in confusing and rambling text, and this change would lead to better results.

In various articles from The New York Times, Associated Press, and The Verge, users report odd encounters with Bing's AI assistant. In addition to getting basic information incorrect, such as pulling out-of-date news articles based on search queries or just responding with inaccurate statements, the program has also given out responses that are combative, and even a bit creepy.
One user, web maker Jon Uleis, shared on Twitter an interaction with Bing's AI tool. What started as a simple question about theater showtimes for Avatar: The Way of Water quickly took a turn when the AI posted a response saying the movie wasn't out yet and that the user was wrong in stating that it was the year 2023 (it is, in fact, was 2023). It is an awkward and amusing back-and-forth to read, yet it does raise concerns about the AI's accuracy in information gathering and the language it uses to interact with users.

Other users had also taken to Bing's AI subreddit to share their stories of what happened when Bing's AI went off the rails. Many of the bad encounters users have had with the AI feel like they come straight out of a comedy satire showing the perils of fast-evolving technology, much like in the vein of Mike Judge's Silicon Valley or Boots Reilly's Sorry to Bother You. Unfortunately, this is the real deal. One reporter from The Washington Post even got responses from the AI that showed that the program had struggles with its own "personality."

As an explainer, the new Bing with AI functionality uses technology from OpenAI, the same company behind the popular ChapGPT tool. In January, Microsoft invested billions into the company to capitalize on the potential AI for supplementing work for the engineering side and improving the user experience for web browsing, programming, and general use for getting information. Much like other AI tools such as the controversial Midjourney or DALL-E, the AI's content is not made within a vacuum but instead pulled from data taken from the internet or fed explicitly into it by programmers. The responses are a high-end predictive text based on prompts and keywords but simulated to create an experience similar to a back-and-forth online conversation.

In the wake of massive layoffs across the tech and online media industries, numerous companies have announced plans or had already started using AI to produce work. With the rapid influence that AI has, there are growing concerns that companies looking to the new technology to lower labor costs and its workforce to maximize profits. This especially evident in the artistic community, with Midjourney and other AI image makers pulling from existing work without permission from artists. For now, AI continues to be a controversial topic, and something that tech industry at large has a keen interest in, and it'll no doubt continue on at a rapid pace.

When the new service made its debut this month, Microsoft stated that the AI was a work in progress and would only sometimes be correct at launch. However, the frequency of these weird interactions is likely not what they had in mind. In many ways, these off-the-rail interactions with Bing AI tend to emulate the more unsavory parts of the internet experience. The issue that Microsoft Bing is facing is one credibility for its new tool, and the recent update that'll lead to shorter messages is a step towards improving the tool -- which is growing in capability by the day.

Alessandro Fillari is a news writer for IGN. You can follow him on Twitter and Instagram @afillari.

Advertisemen 336x280

Read Also:

Related Posts
Disqus Comments