Various facets of artificial intelligence
(AI) application have been discussed for a decade. Recently, a biotech VC fund used AI to make investment decisions. The AI assessed 50 parameters, which were critical for assessing risk factors in the biotech industry. In light of this, let us examine why, how, and to what extent AI could enter the boardroom. The question is not “if” but “when”.
The board plays an important role: Asking crucial questions to the management, providing guidance, and taking decisions. Considering fiduciary duties of directors and legal regime putting more onuses on directors, directors constantly need access to intelligent inputs. Inputs on the internal, external, and competitive environment -- market, productivity, employee trends, etc, that too at the speed of thought. Considering recent unveiling of frauds and corporate governance issues, identifying red-flags based on analysis of facts, such as related party transactions, use of funds, pace of growth, license approval history, contracts with restricted countries or agencies, is equally important.
Usually, executive management carries out background work and presents facts and figures to the board. Through the use of AI, the board could have access to independently assessed and analysed data. It may also enable directors to provide independent value-addition to board discussions. Especially, for independent directors, AI could be an effective tool, as they may have limited industry and technical knowledge.
Various issues arise for consideration, in this regard.
If the board were to rely upon AI, in my view, it cannot be one-size-fits-all. Each industry, sub-industry or each company may develop customised AI that, too, based on objectives sought to be achieved.
It’s a known fact that AI output will be as good or as bad as the data and algorithms on which it is based. If AI is trained on wrong or inaccurate or inappropriate data, the output could be off the mark. Also, with a fast-changing world, some recent developments that have a significant impact on the decision may completely get missed as part of the analysis. The question, therefore, will be who decides the choice of data and algorithms.
The next issue will be transparency about the data relied upon and algorithms applied. There already is a demand for bringing in transparency, especially in the context of bias.
Should the board rely upon AI output for its decision making, can that provide safe harbour to directors from liability? Or to claim safe harbour what additional due diligence would be expected from directors?
Witnessing how various regulators are requiring disclosures about AI and machine learning (ML) deployed by relevant industry, one wonders whether boards will have to make disclosures about their reliance on AI for their decision making.
Overall, it will be interesting to see how the liability and accountability landscape develops.
I want to touch upon one other aspect. Much has been discussed about possible jobs losses because of AI, ethical use of AI, intrusion on privacy because of AI, etc. Boards will have to take a decision on how businesses use AI for their day-to-day operations in an ethical manner.
I know, I have raised more questions that provided answers! At this stage, our endeavour should be -- do crystal ball gazing, and be prepared for the future that is not so distant.
I do not believe that we would want to have a world where individuals completely delegate their thinking abilities to machines. Human evolution may, in fact, stagnate if that were to happen. Hence, I feel there will evolve a world of co-existence where AI is used as a tool and there still will be room for “gut feel”.
Professor Muhammad Yunus once observed: Whether and what extent we want to use AI is still our decision.
The writer is partner at Nishith Desai Associates