5 Things You Must Know About AI Policy




Recent advances in artificial intelligence (AI) and machine learning have triggered new socio-economic challenges and opportunities at the same. We humans are living at the beginning of yet another important transition to in human history. This time the transition is to AI era - where humans would not need to do repetitive tasks (such as customer service, manufacturing, IT administration, etc.) as AI systems could perform those tasks much better and at faster speed. While transition to AI era would improve both efficiency and reduce the associated cost, it is going to pose huge social and economic challenges.

According to an estimate, over next decade 15-50% workforce may be out of job due to AI driven automation of business processes. Although AI economy would be creating new types of job opportunities, but those jobs would require specialized skill set and retraining of out of job workforce. That means, governments and organization would need to have an effective program to retrain and upgrade impacted workforce.

Businesses, organizations and governments all over the world are trying to come with AI policies to maximize AI benefits and minimize its cost and any associated risks. A perfect AI policy is one which equally protects interests of both AI producers (tech giants, banks, insurance companies) and consumers.

Following are 5 important aspects which an effective AI policy must take into account:

1. Unemployment

It is certain that in years to come most economies would see huge unemployments across all industries. AI systems and robots would automate and perform most repetitive business process. The impacted workforce would need to be retrained in order to re-enter job market of AI driven economies. Governments and businesses need to work together to devise a comprehensive and robust AI policy which would not only improve business process but also enable employment in a newly transformed society. Universal Economic Benefits (UEB) - taxing AI business and then paying to citizens; maybe one of the ways to address unemployment caused by AI

2. AI infrastructure

Any government or organization's key interest would be to attract best talent and maximum AI investment share. Right now 40% of AI talent is in United States of America (USA). Governments all over the world need to facilitate research, start state-of-the-art AI education programs and institutes. To attract international talent, governments need to come up with better immigration policies.

3. Standards and Regulations

Since data is the main ingredient of any AI system, there is an utmost need to put robust regulations for data privacy, security and safety. Govts, lawyers and right activists must consult experts from AI community as there is a serious concern about over-regulation of AI. Over-regulation can potentially impact innovation in AI and machine learning.

4. AI bias

AI has a wide range of applications across different sectors of industry. While AI bias is unethical in some fields or use case, it can be very helpful in other applications. For instance, AI bias cannot be permitted in law and justice but it can be very help in medical diagnostics and treatment. A comprehensive framework of AI standards and regulations along with model explainability can help to eliminate bias from AI algorithms.

5. AI Ethics

In an ideal world, AI system or agent is supposed to operate independently without human supervision. In that case, there is a need to define AI ethics - how to design, develop and use AI systems. This component of AI policy also requires rigorous consultation among key stakeholders - lawyers, activists, researcher.

Last but not least, there can be no universal AI policy. Every government, country, business or organization has its own strengths and weaknesses and an AI policy would be set up around them.

Got A Data Science Question?

Ask our experts anything about machine learning, analytics or statistics.