Data ethics

This article is part of a series of articles on data ethics. Read more.

Need to know

  • According to the McKinsey Global Institute’s 2018 report, AI could potentially deliver a projected $13 trillion boost to global GDP by 2030
  • AI’s remit includes multiple areas such as virtual assistants, powering news feeds, facial recognition, computer vision, robotic automation, natural language, IoT and machine learning
  • By using forms of AI such as machine learning, brands are developing algorithms capable of processing vast stores of data at high speed, and making informed decisions in the moment
  • Currently used to improve customer engagement and to increase competitive edge, advanced AI tech is transforming marketing communications; generating detailed customer insight that brands can harness to tailor cross-channel messages, refine product recommendations, and drive better-targeted digital advertising
  • AI has unprecedented capabilities, but with such powerful technology marketers must prioritise responsible and ethical AI
  • AI morality depends entirely on the quality of the data feeding it; and if human partiality is transferred to the data and rules powering algorithms – subconsciously or not – results will be skewed. Given the variability of existing datasets, the chances of this occurring are high
  • Such bias can have serious implications for fairness, inclusion and justice
  • Growing awareness of the potential for AI error has led to greater demand for regulation. Singapore, China and India are among the countries calling for best practices that aim to ensure AI initiatives are human-centric, and machine-based decisions remain both transparent and objective
  • Australia’s Chief Data Scientist Alan Finkel has recommended the creation of a compulsory ‘Turing Certificate’ that AI firms must apply for to prove their trustworthiness
  • Companies must make their AI procedures robust. And to do that, there are four crucial areas they must consider: develop an AI ‘checklist’, obey data laws, minimise machine bias and respect consumers’ privacy choices
  • Stronger data protection rights offer marketers the opportunity to identify and engage with the consumers who really want to hear from them. This is favourable to the current ‘spray and pray’ approach
  • To preserve loyalty and positive brand reputation, brand marketers must optimise AI morality and results by continuously working to maintain a clean, unbiased, unified, and reliable data foundation