Generative AI: Will it change the way we use apps?

What are Google and Microsoft offering?

In Google’s Gmail and Docs, generative AI will help users write documents automatically. For instance, an HR executive can simply ask the AI app to write a welcome email for employees, instead of typing out the document. Similarly, Microsoft has ‘Copilot 365’ for its Microsoft 365 apps, which includes Teams, Outlook, Word and Excel. Here, AI could generate a spreadsheet on command, or even write down an entire article on Word (depending on the topic). Copilot can also match entries on Calendar with emails, and generate quick, helpful pointers that a person should focus on in their meetings.

Are these the only such products?

While Microsoft has already added bots to its Bing search engine and Edge browser, based on AI research firm OpenAI’s GPT-4 model, Google is expected to add a bot called Bard to its flagship search engine. Both firms are making generative AI platforms and models a part of their cloud offerings too, Microsoft Azure and Google Cloud. Since Google and Microsoft are also among the biggest cloud service providers in the world, this means more firms will be able to use these services in order to produce generative AI-based products. Microsoft also offers Github copilot, which helps write computer code.

Who is ahead in the race?

On paper, Microsoft. But that’s because Google’s generative AI products are being tested with limited users, while Microsoft has done larger scale rollouts. That said, the ‘Transformer’ AI model —the underlying system behind all these systems — was invented in 2017 by Google. That means even if Google is behind in terms of public rollouts, it does have considerable experience in this space.

How will the products impact consumers?

By integrating generative AI into their products, Google and Microsoft are increasing the number of people who can access and use such products. In theory, these products are meant to help take the load off humans by taking on the grunt work from day-to-day tasks. For instance, while AI can write a welcome mail for an HR person, the human can put finishing touches to the mail to personalize them. Writers can use these to generate rough drafts of articles, while sales teams can use them to find key trends.

Do we need humans at work at all?

Despite their popularity, the tech is far from accurate right now, prone to giving wrong answers. Both Google and Microsoft’s original demos of these products were found to be giving inaccurate answers. Experts say that while these products can be useful, they aren’t ready to replace humans at work. Since these products learn from the internet or training data which itself may have been inaccurate, they could increase the spread of misinformation online. And who better than humans to check AI-generated info.