Srikanth
Google and Microsoft are actively planning to remove the drudgery from computing, with the help of the launch of next-generation AI tools as add-ons to existing services.
On March 16, Microsoft has made the introduction of an AI-powered system that was named Copilot. Copilot is soon going to get launched to its 365 suite apps. The list of apps is inclusive of Word, Excel, PowerPoint, Outlook, and Teams.
Google published in a blog that the platform is revealing the entire details of its plans to embed AI into its Workspace apps like Docs, Sheets, Slides, Meet and Chat.
All together with this, millions of people are having the usage of these apps on the daily basis. People are bolstering them with AI which can be also helpful in facilitating a major productivity boost.
With the advent of generative AI, AI was mainly used for the categorization and identification of tasks on a previous basis. It accomplishes the tasks like recognizing a number plate using a traffic camera.
Generative AI grants permission to the users the creation of new content, by applying deep-learning algorithms to big data. ChatGPT and DALL-E have already accepted the storm of global transformation.
But now this is the time when Microsoft and Google have found a more concrete way to bring generative AI into offices, classrooms, and many more places.
There are various other generative AI tools, like them, Copilot and Workspace AI are also built on large language models (LLM) trained on massive amounts of data. With the help of this training, the systems have “learned” the variety of rules and patterns that can be applied to new content and contexts.
Microsoft’s Copilot is being trialled on 20 customers. In the coming months, Copilot will define the details of availability and pricing.
Copilot will be integrated across apps to help expedite the tedious or repetitive tasks. The launch of Copilot will be helpful for users to write, edit, and summarize Word documents. It has the capability to convert ideas or summaries into full PowerPoint presentations. It can also identify the data trends in Excel and can also quickly create visualizations. If users have the desire to “synthesize and manage” the Outlook inbox, then Copilot is the perfect choice for them. It provides real-time summaries of Teams meetings. Bring together data from across documents, presentations, emails, and calendar.
It has been assumed and expected that all these tasks can be run effectively by Copilot. Copilot will be a massive upgrade from Microsoft’s original Office Assistant, Clippy.
Google’s Workspace AI will offer similar capabilities for the payers of subscribers.
Microsoft has also revealed the description of Copilot as “a sophisticated processing and orchestration engine. The team of the company is working on combining the power of LLMs, including GPT-4.
Google’s Workspace AI is built on PaLM that stands for Pathways Language Model, which was trained on a combination of books, Wikipedia articles, news articles, source codes, filtered webpages, and social media conversations.
Both systems are integrated with existing cloud infrastructure. This simply states that all the data they are applied to will be online and stored on the server of the company.
There is a huge requirement for complete access to the tools for the relevant content in order to facilitate contextualized responses. For instance, Copilot can’t distil a 16-page Word document into one page of bullet points without making the analysis of the complete text.
Now generally the question arises of whether will users be informed to train the underlying models. With respect to this, Microsoft has said that Copilot’s large language models are not trained on customer content or on individual prompts. On the other side of the page, Google mentioned that the private data is kept private, and not used in the broader foundation model training corpus.
There is the possibility of availability of temptation to train such tools on real customer-specific data in the future. But for now, seems to be explicitly excluded.
According to the usability concerns, there are various people noted following ChatGPT’s release, text-based generative AI tools are prone to algorithmic bias. These concerns will be helpful in extending the new tools from Google and Microsoft.
Bing chatbot which is owned by Microsoft also runs on GPT-4 which is capable enough to create outrageous claims.
Biasness takes place when large volumes of data are processed without the selection and understanding of appropriate training data, and without proper oversight of training processes.
Just for example, there is a lot of content that is available online written in English which is likely the main language spoken by the people developing AI tools. This underlying bias can influence the writing style and language constructs.
Basically, it becomes hard to determine how issues of bias might present in Copilot or Workspace AI.
Regarding security concerns, one of the major vulnerabilities in Microsoft’s and Google’s AI tools is that they could make it much easier for cybercriminals to bleed victims dry.
Whereas before they trawl a criminal through hundreds of files or emails so that they can identify the specific data. With this, the criminals can make the use of AI-assisted features to collate and extract what they particularly require promptly.
Also, since there’s no indication of offline versions, anyone who has the desire to use these systems will have to upload the relevant content on an online basis. Uploading data over the online platform is at greater risk of being breached than data stored only on your computer or phone.
From a perspective of privacy, it’s not particularly inspiring several avenues through which the biggest corporations in the world can collect and synthesize our data.
