Hitachi Ltd. says it has developed an “artificial intelligence” system that draws on a massive range of data sources, such as millions of news articles, and can provide a reasoned response to hot-button topics.
The firm said Wednesday the joint project with researchers at Tohoku University had produced big-data software that backs up its decisions with facts, background and context.
For instance, when asked whether casinos should be banned or not, the software replies: “Without the support of the local residents . . . the ban on casino business should be maintained.”
The system currently uses English, but the company said it plans to introduce Japanese by the end of the year.
It said examples of where it might be used include deciding medical treatment based on a patient’s clinical records or making business decisions at firms that handle large amounts of data.
The system currently has access to 9.7 million news articles and reports, indexed by what it describes as a web of 250 million correlations.
One source it draws on is a database of health, safety and economics information. It was compiled by an organization whose expertise includes forensics. A section of this database, developed by a Hitachi subsidiary, was tested at an international forum, where it was found to be more accurate in determining a health problem and locating a disease than existing systems.
Hitachi also developed a database that places the value of an action into two categories — positive and negative.
For example, physical activity is regarded as a positive aspect of health, while disease or obesity is recognized as negative. They are further prioritized according to a term’s level of importance and based on frequency of usage.
Hitachi says if users require an urgent answer the system can cut short its searching.
With further development, Hitachi says users may be able to conduct logical debate with the software.
In a time of both misinformation and too much information, quality journalism is more crucial than ever.
By subscribing, you can help us get the story right.