Microsoft now wants to add ChatGPT to Word, email

Microsoft is reportedly in discussions to bring OpenAI’s AI writer ChatGPT to Microsoft Office as part of an effort to modernise the office software suite.

The news, courtesy of The Information (opens in new tab) via two insiders at Microsoft, comes after TechRadar Pro recently reported that the company is looking to bring OpenAI’s ChatGPT to its Bing search engine for web browsers.

Plans to bring ChatGPT’s text-generation capabilities to “Word, PowerPoint, Outlook and other apps” follow another stalled attempt to implement AI tools that use OpenAI’s machine-learned language models into Office to generate emails and documents.

 Microsoft, ChatGPT, and the future

According to the insiders, Microsoft wants to leverage ChatGPT to “provide more useful search results when Outlook email customers look for information in their inboxes.”

“For instance, GPT has the ability to figure out what emails the customer might be searching for even if they don’t type the exact keywords that are in the relevant emails.”

Moving forward, it’s expected that Microsoft will want ChatGPT to write e-mail replies, rewrite documents for clarity, and produce entire documents according to prompts, similar to how ChatGPT currently works.

 ChatGPT has created a culture of fear in its recent rise to prominence, as the world speculates what it might be capable of, despite limited evidence that it’s anywhere near ready. 

Google executives are concerned about its inaccuracy, and educators are worried it will break an already broken system, but even Sam Altman, the CEO of OpenAI, says that ChatGPT is unreliable (opens in new tab), and at most a proof-of-concept at this stage.

It’s entirely possible that Microsoft may have big plans in mind for ChatGPT and other products driven by OpenAI’s GPT language model, but as yet, there’s no fixed estimate for when they might become a reality.

There’s a reason why ChatGPT is currently marked as a “free research preview”, and peppered with warnings that it “may occasionally generate incorrect information”, “harmful instructions or biased content”. It’s not there yet.

And until it is, instead of chasing the tails of insiders and how companies may or may not end up leveraging nascent technology, maybe we would all benefit from doing something else with our lives until ChatGPT is at the point not just where it can impress us, but that it’s also readily applicable to different aspects of our lives.