What is Artificial Intelligence?
Simply put, Artificial Intelligence (AI) is the ability some machines have to “learn from experience” and improve their performance without needing a human to reprogram them. “Learn from experience” is between quotes here because strictly speaking, experience and learning can only happen to conscious beings, not machines — so what exactly do we mean when using those words to talk about computer programs? How does software “learn”?
Computer systems “learn” by performing three basic steps:
- First, they amass and store huge quantities of data.
- Then, they analyze the data to find patterns and trends in it.
- And third, they use those patterns and trends to adapt their reactions to specific inputs.
Each time there’s a new input, the system will react differently, because it has new data on which to base its reactions. To people, it feels as if the system were “learning”.
Basic examples of Artificial Intelligence
As an example, think of YouTube video recommendations. When you watch a video, the YouTube system records that information, then activates algorithms that compare it to an enormous pool of data from other users. The algorithms analyze the patterns and trends of other viewers’ activity — people of your sex, within your age range, in your geographical region — and based on that, selects which videos you are most likely to be interested in watching next.
Now suppose you watch one of the recommended videos and mark it as “disliked”. The system will store that information as well, so next time you watch that first video, the disliked video will not show again in the recommendations. The algorithms have “learned” something about your tastes, and their reaction has adapted to it.
This is a basic example of AI at work. Other common examples of AI are:
- computer vision
- automatic customer service on websites
- speech recognition
How does AI in data rooms work?
When it comes to virtual data rooms (VDRs), there are several features that use AI.
Most common AI features in VDRs:
- Automatic indexing. When you upload a batch of documents onto the VDR, the system indexes them according to the keywords they include. If it indexes a document according to an irrelevant keyword — or fails to index a document under a pertinent keyword — you can correct that, and the system will register your input to refine its indexing.
- File naming suggestions. Based on the name you enter for the first file in a batch, and on the contents of the other files, the system will suggest file names for each file in the batch. After you accept or refuse the recommendations, this feature will save your preferences to improve naming suggestions on the next upload.
- Document search, analysis and summaries. These tools will conduct searches according to keywords or key phrases you enter and find all pertinent documents. They will then analyze the document contents and summarize them for you.
- Document review. Based on the keywords and phrases you enter, this tool will review document contents to locate any errors or potential issues. The best review tools will do this in many different languages.
Advantages of AI-powered data rooms
The main advantages of using AI-powered data rooms are:
- saving time by performing automatically tasks that would otherwise be done by people
- reducing cost by spending less time on due diligence processes
- minimizing or altogether eliminating human error in tasks that can be safely automated
- increasing the accuracy of VDR tools over time