Law Abiding AI Applications
Asimov's 2nd law states 'A robot must obey orders given it by human beings except where such orders would conflict with the First Law.'
By Samik Kumar – Chief Technical Officer, Level
In the modern context, Artificial Intelligence (AI), bots and robots are part of everyday discussions, so today it becomes more relevant to understand whom the AI application would obey. If it was in my house there will be no room for ambiguity, it would be the boss of the house (my better half to avoid any doubts that you may have). But when we put this in the context of AI, cognitive and robotic applications used by organisations and business functions, things get blurred. It becomes ambiguous and potentially emotive. Should the application listen and obey business functions, or should it listen and obey the IT department, in the same way as they control and manage other business applications?
AI as part of the team
Organisations expect AI applications to work as part of the business, and like other members of the team, they are expected to be managed by the respective business function. Like other employees, these applications will be trained to learn new skills or to enhance their existing skill set e.g. regulatory changes. When they do not interpret policy correctly, business managers should be able to correct them. Only when they stop working or make mistakes that their managers within the business cannot correct, specialists are required from the IT department who can correct these applications and get them back to work.
For this to happen, all three collaborators involved; software companies, business functions and IT departments, must change their approach.
Software Companies: they must build applications that are easy to configure, use and manage with negligible technical skill. Applications must be such that they can be configured and managed using natural language, following simple, easy to understand instructions. They need to provide the necessary tools the business needs to enhance their productivity and efficiency to maximise the use of cognitive applications.
Business Functions: before they take ownership of applications, they must ensure their business processes and policies are documented, and business rules clearly defined for them to configure these applications. To work and think like humans the cognitive applications need to understand all policies and business rules and this knowledge must be transferred from the subject matter experts in the business to the applications.
IT Departments: IT departments must redefine their role and become an enabler and provider of a secure ecosphere for these applications to work and grow in. They must become the highly skilled task force that supports the business when they need the 'A' team to intervene. With the growth of cloud infrastructure and SaaS applications, it is becoming easier for IT departments to embrace this change.
Find out how Level can help your business be more productive with the use of AI.