Google Crisis Information Systems: AI to improve detection of crisis-related searches

Currently, when people search for sensitive information — such as suicide, abuse, or other topics — Google will display the contact information for the relevant national hotlines above its search results. However, the company explains that people in crises may search in various ways, and it’s not always apparent to a search engine that they’re in need, even if it would raise red flags if a human saw their search queries. 

Google claims that with machine learning and the latest improvements to its AI model called MUM (Multitask Unified Model), it will be able to automatically and more accurately detect a broader range of personal crisis searches due to MUM’s ability to better understand the intent behind people’s questions and queries.

At its Search On event, the company unveiled its aim to rebuild Search using AI technology, although it had not yet addressed this particular use case. Google then concentrated on how MUM’s improved comprehension of user intent may assist online searchers in gaining a more profound knowledge of the subject they are studying and guide users along new search avenues.

Now, MUM will be employed to comprehend better the subjects that someone in need might look up, which aren’t always as evident as typing in a natural cry for assistance.

Beyond the fires

Google has plans to expand and develop the program over time to include data like traffic conditions and road closures in addition to fires.

According to Google sources, the Crisis Map application could be updated to identify Australian threats from tropical storms, cyclones, floods, and landslides, as has been done in other nations like Canada, Colombia, India, Japan, and Taiwan. Currently, the Crisis Map application focuses on bushfires and mirrors the map provided by the NSW Rural Fire Service.

The Crisis Map project is slightly different from Google’s conventional data collection methods, and a shift towards greater reliance on data feeds from third parties, including rural fire departments, emergency agencies, and the general public.

While the new application may have the potential to improve upon existing websites and online services, its real success will depend on such crucial elements as:

• the information provided’s usefulness
• the applicability of its contents
• the capacity of the application to establish itself as a reliable information source.

Therefore, it appears that the availability of the most recent and up-to-date information may be more critical to the quality of information offered by the program than simply how rapidly the Crisis Map application can synthesize data gathered from various sources.

Checkout our new blog on GOOGLE BERT Algorithm!

Real life in real-time

Bushfires are the most dynamic and quickly spreading natural occurrences outside earthquakes and tsunamis.

Information regarding bushfires or other risks must, in theory, and in the majority of circumstances, be gathered and delivered from blaze zones in real-time mode utilizing real-time data feeds to be truly useful to the majority of users who reside in or travel through vital areas. They should offer any available information on the following:

• the direction of the fire spread 
• its magnitude
• its velocity and acceleration
• its shape and other relevant data.

The efficient coordination and implementation of these aspects in the near future are still impossible. However, it is not impossible given the rapid development of satellite communications, smartphone technologies, wireless sensor networks, the Internet of Things, as well as signal and image processing techniques.

Therefore, the new Google effort might initially serve an educational and awareness-raising purpose. The most up-to-date or essential information on bushfires could not always be accessible, either as a result of the environment surrounding them changing very quickly or even as a result of internet infrastructure breaking down due to fires.