This submit is the primary in a Constructing AI Responsibly sequence, which explores prime considerations with deploying AI and the way Microsoft is addressing them with its Accountable AI practices and instruments.
Technically, hallucinations are “ungrounded” content material, which suggests a mannequin has modified the info it’s been given or added further data not contained in it.
There are occasions when hallucinations are helpful, like when customers need AI to create a science fiction story or present unconventional concepts on every little thing from structure to coding. However many organizations constructing AI assistants want them to ship dependable, grounded data in situations like medical summarization and schooling, the place accuracy is important.
That’s why Microsoft has created a complete array of instruments to assist handle ungroundedness primarily based on experience from creating its personal AI merchandise like Microsoft Copilot.
Firm engineers spent months grounding Copilot’s mannequin with Bing search knowledge by retrieval augmented era, a way that provides further information to a mannequin with out having to retrain it. Bing’s solutions, index and rating knowledge assist Copilot ship extra correct and related responses, together with citations that permit customers to search for and confirm data.
“The mannequin is superb at reasoning over data, however we don’t suppose it ought to be the supply of the reply,” says Fowl. “We predict knowledge ought to be the supply of the reply, so step one for us in fixing the issue was to carry contemporary, high-quality, correct knowledge to the mannequin.”
