Last year in February, IBM was one of the signatories to the Rome Call for AI Ethics. It was supported by Pope Francis and the agreement was signed in the presence of the president of the Pontifical Academy for Life, Archbishop Vincenzo Paglia. Two things stand out: AI is serious enough for even the Vatican to be deeply involved in, and the kind of goodwill that IBM has earned across the world.

The CTO of IBM India, Subram Natarajan was in conversation with Jibu Elias of INDIAai. Subram Natarajan is the Chief Technology Officer of IBM India/South Asia and leads IBM's platform conversations with key customers around cloud and cognitive software, to help them address their IT and business challenges. And in a fireside chat, the duo touched upon some thought-provoking ideas behind the guiding principles of AI ethics. This column is an attempt to capture some of the leading thoughts centred on TRUST & TRANSPARENCY. 

Even before the pandemic, leaders proclaimed that AI adoption would soon be a hygiene factor for most companies or at least those which were going to be in contention. Covid cemented this argument and accelerated digital transformation journeys many times over. AI adoptions have now started to become more mainstream and companies are investing heavily in building competence around Data Science, ML, and allied areas. IBM is helping many companies by providing the platform including application tools and advising them on the right methodology to get there. 

AI should augment human intelligence and not replace it – this belief forms the cornerstone of IBM’s approach. AI is, after all, a very powerful technology that can take advantage of mountains of data which is not possible through legacy systems. Intelligent Automation drives operational efficiency in a way that rapidly enhances value-creation and pushes practitioners up the value chain not quite seen earlier. The challenge today is about talent shortage and the ecosystem participants (industry, academia, and the government) are all up in arms to reduce this gap. 

If AI adoption is to scale, then the model has to be open & trusted. More than 60% of the CEOs surveyed are still apprehensive about compliance and whether these models can meet those standards. For scalability, this is a key ask.

Principles that IBM swear by

• AI should augment humans and not replace them. AI should make our jobs a lot better and it’s not only for a few people but many.
• Data & insights belong to the creator. In IBM’s case, it’s their clients.  
• All AI systems must be transparent and explainable. Tech companies will have to be very clear about this aspect. E.g., What kind of data was used and the algorithms that led to the recommendations. There has to be conclusive evidence towards this and if that’s not the case then the model shouldn’t be out there. It’s non-negotiable.  

Earning a high level of trust is not easy and challenges are faced due to certain factors:

• When and for what purpose AI is being used – remains unclear in certain cases
• Expertise in handling the data – the governance & policy angles also kick in.
• Are the clients' data & data-driven insights protected through adequate measures?