#CIOChat: Ethics in AI; Will enterprise take note?

Tweetchat participants: 

Kiran Belsekar, SVP, Head – IT, ENAM AMC ; @kit_kiran

Mukesh Mehta, CTO, B&K Securities; @connect_muks

Rajiv Sharaf, Sr Executive VP & CIO, Reliance Infrastructure ; @RAJIVHARAF

Moderator: Shantheri Mallaya, Executive Editor, IDG India; @shaanemaan 

IDG

According to Gartner, from now until the year 2022, 85 percent of artificial intelligence (AI) projects will deliver inaccurate outcomes due to bias in data, algorithms, or the teams responsible for handling them. While we are talking about AI and its various levels of implementation, there is a need to step back and analyze  – did we miss out on incorporating ethics into it?

The European Commission (EC) released one of the world’s first government-led set of guidelines on developing and implementing ethics in artificial intelligence. These guidelines will not only serve as a pioneering roadmap for organizations poised to leverage the full potential of AI, but who may be apprehensive due to potential issues surrounding trust and transparency.   In this context, CIO India hosted a tweetchat on April 25, 2019, and asked leading CIOs in the country if there was a need for enterprise to take note of implementing ethical AI. 

Defining ethical AI 

Defining ethical AI was the first talking point of the discussion. And every CIO had their way of looking at it. For example, Mukesh Mehta, CTO of B& K Securities says, “Ethical AI should ensure goals and benefits of AI are not restricted to the creator or any particular community. Though, it is by nature that AI strives to pursue the primary goal (business progress, wealth accumulation, etc).” 

While Kiran Belsekar, head-IT, ENAM AMC, believes that ethical AI is the any technology that can answer – is it moral, it is safe and is it right positively. Additionally, the technology should be developed in accordance with international laws. 

Dilemmas in AI

Talking about the dilemmas of AI, CIO’s said forms of AI corruption include misinformation, anti-social happenings being reported as norms, antithesis and anti-sentiments.  

A2.1 Open access to knowledge available in different mediums is essential for AI’s success. Unfortunately, our knowledge sources are so skewed & dangerous it is becoming impossible to keep AI from getting corrupted. #CIOChat

— Mukesh Mehta (@ConnectMuks) April 25, 2019 

According to Belsekar, the biggest dilemma is the trust one can choose to put on it. “Though AI is capable of a speed and capacity of processing that’s far beyond that of humans, it cannot always be trusted to be fair and neutral,” he said. 

CIOs also expressed the need for transparency in AI powered machines, and that there could be two ways to achieve it. 

A3.2 It’s essential that creators of an AI system have the right intentions. Auditing the intentions of the AI system being built is essential in the beginning itself. #CIOChat

— Mukesh Mehta (@ConnectMuks) April 25, 2019

“Responsible AI is more than the ticking of some ethical `boxes’ or development of some add-on features in AI systems. It requires participation and commitment of all stakeholders, and active inclusion of all of society. This means training, regulation and awareness,” says Belsekar. 

During the conversation tech leaders opined that AI and ethics woven into AI powered systems are the collective responsibility of communities, and using AI for the larger good in society and community development is one way ahead. 

Enterprise, innovators and the industry should get over the fear of bias in AI by marrying business and technology. Then they will be able to arrive at a 360-degree vision that will consider compliance and ethics, and also be viable, sustainable and scalable.   

CIOChat is conducted every fortnight around a topic that is trending and relevant to the CIO community in India. It is hosted on @CIOin and moderated by our Executive Editor, Shantheri Mallaya. In case you want to join the chat or get more details on the upcoming chat you can mail our Executive Editor shantheri_mallaya@idgindia.com.