So far, 2023 has been the year of AI. From Bard to ChatGPT, the news cycle in the first quarter of the year has been dominated by stories celebrating both the great and controversial aspects of artificial intelligence.
Well, last week was no different, and we saw major launches from both Microsoft and Baidu… with varying degrees of success. Here, we recap both stories.
Microsoft announces Copilot, the AI-powered future of Office documents
Last week, Microsoft announced an AI-powered Copilot for its Microsoft 365 apps. Designed to assist people with generating documents, emails, presentations and more, Copilot will be powered by GPT-4 from OpenAI.
Copilot will sit alongside Microsoft 365 apps much like the beloved Clippy paperclip used to. It will appear in the sidebar and users will be able to summon it in order to generate text in documents, create PowerPoint presentations based on Word documents, or even help use features like PivotTables in Excel.
Described as a “whole new way of working” by Microsoft 365 head Jared Spataro, users will also be able to use Copilot to provide information on an upcoming Microsoft Teams meeting and prepare people with updates on related projects.
Plus, Copilot can also be summoned throughout Microsoft’s Office apps and can be used in Word to draft documents based on other files. The AI-generated text can then be freely edited and adapted. As Copilot is essentially a chatbot, users can also ask it to create a 10-slide PowerPoint presentation based on a Word document or analyse or format Excel data.
Similarly, Copilot will also exist in Outlook. This means that users can spend less time deleting emails and responding. Email threads can be summarised, and Copilot can even create draft responses with toggles to adapt the tone or length of an email.
However, although it has the power to transform the way Office users work, Spataro has stressed that Copilot will not always be correct. During Microsoft’s AI event last week, he told the audience that “Sometimes Copilot will get it right, other times it will be usefully wrong, giving you an idea that’s not perfect but still gives you a head start.”
Microsoft says it’s testing Copilot with 20 customers right now and will be expanding the preview in the coming months.
While Microsoft is moving quickly with its AI-powered vision for Office apps, concerns around the speed of innovation and the accuracy of its AI models will remain; particularly when Microsoft 365 users are providing the AI with business data.
Baidu’s ChatGPT offer falls short of expectations
Meanwhile, last week a Chinese ChatGPT rival from search engine firm Baidu failed to impress. After the initial launch of its ChatGPT-like artificial intelligence software known as Ernie Bot, Baidu’s shares fell by as much as 10%.
While Chinese tech companies such as Alibaba and JD.com have also announced projects to rival ChatGPT, Baidu’s Ernie Bot fell well short of expectations. Due to fears over performance, the company’s co-founder and chief executive, Robin Li, only showed a pre-recorded demonstration of the software’s capabilities, rather than a live interaction.
Li himself admitted that the technology was still flawed but was being released to meet huge customer demand. “Our expectations for Ernie Bot are close to ChatGPT, even GPT-4,” Li said.
Why does this matter?
Both of these releases show us that AI is here to stay. In fact, it’s clear from both of these unveilings that companies are now rushing to release AI chatbots in order to meet customer demand. However, it’s worrying that this race to corner the AI market is now resulting in the release of chatbots that are not quite ready for the market.
As a result, we think that users should be wary when using AI. Although the technology has seemingly limitless potential, it’s worrisome that companies are releasing AI-powered chatbots that are giving inaccurate replies. In fact, it’s particularly worrying that Microsoft is describing its own AI chatbot as being “usefully wrong”, which is oxymoronic at best. While AI provides many advantages that must be embraced, users must be aware that the technology is not perfect and rigorous checks must remain in place.
Author spike.digital