Microsoft CEO Satya Nadella speaks at the company’s Ignite Spotlight event in Seoul on November 15, 2022.
Seong Joon Cho | Bloomberg | Getty Images
Thanks to recent advancements in artificial intelligence, new tools like ChatGPT are wowing consumers with their ability to create compelling writing based on people’s queries and prompts.
While these AI-powered tools have gotten much better at producing creative and sometimes humorous responses, they often include inaccurate information.
For example, in February, when Microsoft launched its Bing chat tool, built using GPT-4 technology created by Microsoft-backed OpenAI, people noticed that the tool was give wrong answers during a demonstration related to reporting financial results. Like other AI language tools, including similar software from GoogleBing’s chat feature can occasionally presenting false facts that users might believe to be the basic truth, a phenomenon researchers call “hallucination.”
These issues with facts haven’t slowed the AI race between the two tech giants.
Tuesday, Google announcement it brought AI-powered chat technology to Gmail and Google Docs, allowing it to help compose emails or documents. On Thursday, Microsoft announced that its popular business apps such as Word and Excel will soon be paired with ChatGPT-like technology. nicknamed co-pilot.
But this time, Microsoft presents the technology as “usefully fake”.
In an online presentation about Copilot’s new features, Microsoft executives pointed to the software’s tendency to produce inaccurate responses, but touted it as something that could be useful. As long as people realize that Copilot’s answers might be sloppy with the facts, they can correct inaccuracies and send their emails faster or finish their presentation slides.
For example, if someone wants to create an email wishing a family member a happy birthday, Copilot can still be useful even if it has the wrong date of birth. In Microsoft’s view, the simple fact that the tool generates text saves a person’s time and is therefore useful. People just have to be very careful and make sure the text is free of errors.
Researchers might disagree.
Indeed, some technologists like Noah Giansiracusa and Gary Marcus expressed concerns that people may be placing too much faith in modern AI, taking advice tools like ChatGPT to heart when they ask questions about health, finances, and other high-stakes topics.
“ChatGPT’s anti-toxicity barriers are easily circumvented by those who want to use it for harm and as we saw earlier this week, all the new search engines keep hallucinating“, wrote the two in a recent Time opinion piece. “But once we get past the jitters of opening day, what will really matter is if one of the big players build an artificial intelligence that we can really trust.”
It’s unclear how reliable Copilot will be in practice.
Microsoft chief scientist and technical colleague Jaime Teevan said that when Copilot “mistakes or is biased or misused”, Microsoft has “mitigating measures in place”. Additionally, Microsoft will test the software with just 20 enterprise customers initially so it can find out how it works in the real world, she explained.
“We’re going to make mistakes, but when we do, we’ll fix them quickly,” Teevan said.
The business stakes are too high for Microsoft to ignore the excitement around generative AI technologies like ChatGPT. The challenge for the company will be to integrate this technology so that it does not create public distrust of the software or lead to major public relations disasters.
“I’ve studied AI for decades and I feel this tremendous sense of responsibility with this powerful new tool,” Teevan said. “We have a responsibility to get it into people’s hands and do it the right way.”
