Skip to main content

AIU@Work: What is Explain-ability?

There is no question that installed AI has a set of risk factors all its own. Fears of disruption, threats of infiltration and security risks cloud the safety and dependability of AI systems. System failures can add to the complexity of AI operations. Data can be lost. Deep learning uses data in image and speech recognition. The data involved and how it is used is really not understood by most humans. The arrival of AI on a much larger scale has also opened questions of privacy. Although the robots are here, very little has been put into place as far as legal and regulatory preparations. Europe currently has the General Data Protection Regulation (GDPR). The U.S., China, India and other countries don’t have much data protection.

There is no question that cloud-based data has changed the way we connect the world. Life on the cloud for data has gained in popularity. Many companies and organizations are concerned, however, about risks vs. benefits. Explaining how computers behave can be complex. It is a difficult endeavor even if people are willing to do it. Regulators are challenged by deep learning. They tend to look at AI models like black holes. The results can be accurate but “how we got there” remains a mystery. Companies are pouring huge amounts of money into explain-ability. We have had Facebook disasters and fake news in the 2016 U.S. elections. Everyone realizes there is an issue with false information and AI. Misuse of personal data can affect important life decisions. It affects credit worthiness, crime detection and bias.

Biased data can generate biased results. One study showed that Google showed ads for high paying jobs more to men than to women. These kinds of issues have challenged us to dig deep and consider what regulatory and legal measures are needed. We need regulation so that with AI we don’t run the risk of discriminatory and offensive results. The arrival of AI is not without challenges.

#AI@Work, #WFH, #Virtual Touchpoints, #ThePajamaEffect, #The Visual Connection, #BobbeGB, #BobbeBaggio, #Touchpoints, #Remote Workplace, #WorkFromHome, #PJEffect, #LinkedInNewsLive