Two broad themes stood out at AIS 2018

• Repositioning AI’s end game from replicating human brain to augmenting it

• And building AI for the greater good

Speakers discussed numerous promising trends

• Huge advances in the past two years across applications

  • Vision most popular example

  • Excitement regarding progress in deep learning

  • And auto-machine learning (ML) and increasing ROI

• AI is operating at scale

• Recognition of the need to address public/press (bots will replace humans…) andethical/regulatory concerns

  • IBM announced its explainable AI dashboard which monitors “fairness” aka bias

  • Genpact has impressive click-thru feature to drill down to atomic level

  • Google recently released AI Ethics

  • Several presenters recognized the important role subject matter experts will continue to play in not only ML fine tuning but also ensuring fairness

• Bridging gap between data scientists and end-users/non-data scientist

• New models to get more input from consumers – enabled by smartphones, sensors, and related and deliver value add

  • Top use case includes healthcare (e.g., encourage best practice post-op with physical therapy)

  • A standout example in banking:Capital One’s interactive card fraud alert
    (a Celent Model Bank 2018 winner for its Fraud Unlock)

    Most AI active industries remain the same:

• Banking – fraud mitigation, KYC facilitation; customer service

• Insurance - claims

• Healthcare/pharma – patient care, empirical medicine, diagnosis

• Retail, digital services – recommendations, call centers

  To learn more about AI in banking, please see my report, From Moonshot to Reality: AI in Action, June 2018.