The proliferation of increasingly complex financial regulation has inevitably inspired technological responses - the so-called regtech of aggregated data, machine learning and automated decision-making.The implications of such an environment was explored via an expert panel discussion at ASIC's annual forum yesterday.Grace Brasington, IBM vice president of global banking and financial markets, said that with the proliferation of data, there were no simple answers: "Ninety per cent of the world's data has come to us in the last two years - and if you wonder at that, think about emails, think about chats, think about blogs - and 80 per cent of that data is unstructured." "Our view is that, to the extent you can merge what the regulatory system requires with technology, then that's a great benefit as we can look at things in a much more holistic way."This was the thinking behind IBM's acquisition of Promontory, a risk management and regulatory compliance consulting firm, about six months ago.The new hybrid operation, using in-depth knowledge of rules and regulations with the massive capacity for finding patterns that IBM's Watson can bring to bear. IBM has also since hired Jeffrey Carmichael, former head of the Australian Prudential Regulation Authority, to run its Australasian business.This need for machine learning and other tech help has not eluded Rebecca Lim, group general counsel and chief compliance officer at Westpac.Regtech, Lim said gave her and her colleagues "the ability to see more clearly across the enterprise and to scale more effectively, and to streamline the way we do business."By way of example, Lim estimated that in 2015 there were over 1800 Australian federal acts and over 18,000 pieces of subordinate regulations that she had to deal with. This explosion of legislation and regulation, and the need translate all legislation into plain English and then to embed it into the organisation was "quite herculean", Lim said.Estimates are that at the current rate of progress there will be 300 million pages of financial services legislation globally by the end of 2020.The third panel member was Shonali Krishnaswamy, chief technology officer of AIDA Technologies in Singapore. Drawing on her five years leading ASTAR (Agency for Science, Technology and Research), which included a long-running machine learning project with DPS, one of the largest banks in Singapore, Krishnaswamy explained that the starting point was building rules and rule engines, but humans were needed to work out how these cut in.At its most sophisticated, pattern recognition can lead to predictive audits, extrapolating risks and warning before things go wrong.However, Krishnaswamy cautioned about the need for expert interpretation of any results. "No artificial intelligence system is going to give 100 per cent accuracy, more like 85 per cent," she said.Her conclusion - one that should please nervous compliance professionals - was clear cut: "AI should be augmented by human intelligence. It is all about making human decision makers focus on what is important."There cannot be machines making the decision but the aggregation of information that is the value proposition."