How ChatGPT in Microsoft Office could change the workplace | The AI Beat

Over the weekend, The Information reported that Microsoft is looking to add OpenAI’s chatbot technology — currently ChatGPT, soon to be GPT-4 — to its Office suite of productivity technologies, including Word, Outlook and PowerPoint.

I immediately wondered: How would these apps-on-steroids, used by billions of companies globally, change how we work? Especially once Google gets fully in the game, integrating its own generative AI capabilities into Google Workspace? Would AI become as mundane in our day-to-day work lives as the humble spreadsheet?

Far more than a new Clippy

News of the plans came only a few days after word spread that Microsoft, which invested $1 billion into OpenAI in 2019, was planning to embed ChatGPT into its search engine Bing. But this report led those of an earlier tech generation to immediately giggle. Why? One word: Clippy.

Clippy, Microsoft’s user interface agent that came bundled with Microsoft Office in 1997 and was personally launched by Bill Gates, was a hopped-up, big-eyed paperclip who popped up to say things like, “It looks like you’re writing a letter. Would you like help?”

Clippy was mostly loathed and mocked for his invasive pop-ups. Time even named Clippy one of the worst inventions of all time. Clippy had disappeared entirely by 2007, although he was resurrected as a cultural icon as a retro sticker pack in Teams in 2021.

ChatGPT, of course, would be far more than a new Clippy — it could potentially do everything from generate text based on simple natural language prompts and suggest responses to emails to analyze data in Excel and translate text.

Tech investor Puneet Kumar called the possibility “crazy powerful” in a tweet, adding that it would “further deepen” Microsoft’s moats in enterprise office tech:

ChatGPT and similar models not ready for prime time

The Information pointed out that Microsoft Word already uses in-house AI tools, including Turing’s Smart Find feature and At a Glance, which summarizes Word documents. And it has “already quietly incorporated GPT into Word in minor ways,” including in its autocomplete feature.

But to implement OpenAI’s ChatGPT or, soon enough, GPT-4, there will be plenty of hurdles to overcome.

For one, ChatGPT has a serious accuracy problem, one that is exacerbated by its tendency to sound plausible even when it is dead-wrong. Even OpenAI CEO Sam Altman has admitted the risks. That makes corporate document creation or advanced workplace integration a no-go at the moment.

Privacy is also an obstacle. How would Microsoft preserve the privacy of corporate data? It’s hard to imagine a large law firm or financial services company that uses Microsoft Office all day long getting help from ChatGPT right now.

Office work will likely change for good

Still, the opportunity to significantly power-charge the day-to-day text output of the average enterprise — emails, presentations, reports — is too tantalizing to ignore. As some guy named Kevin tweeted:

But it may be some time before it’s clear both how Microsoft and Google can make generative AI tools work for business at scale, as well as how enterprises can deal with what employees create.

Forrester analyst Rowan Curran said there are “a lot of open questions about what guardrails and controls enterprises put both on how these tools are allowed to be adopted, and how they can be used once they are adopted.”

For example, he told VentureBeat by email, “If I have a text generator on my phone that I used to draft a work email or outline a blog post and then I publish that as part of my job – does my employer need to be concerned about, or at least aware of that?”

So much about our digital office life — think PDFs, spreadsheets, smartphones, cloud, digital signatures — have become work ho-hums over the past two decades. Whether it’s ChatGPT in 2023 or not, it seems likely that advances in generative AI are on the path to transforming the workplace as we know it.