Code of AI ethics in marketing

anch.AI stives to uphold human ethical values in our use of AI in marketing. We will uphold data privacy and unwanted influencing of the individual by being transparent in data collection and allowing users to opt-out of any marketing activity. We will uphold accountability by always chose for transparency in our marketing technology stack and describing its purpose for our users. We will counteract bias and impact on societal values by monitoring outcomes and adjust input in our marketing technology.



Guidelines for a golden standard in AI ethics for marketing

Organization

  • Involve management as well a cross-functional teams to set ethical norms
  • Ethical training for employees and vendors
  • Have open R&D
  • Always have a human in the loop when designing marketing stacks containing AI

Support

  • Always ask “why?” and “what is the goal?” of any AI purpose
  • Identify and involve key stakeholders
  • Provide non-experts with enough input so they can contribute
  • Assess how ethical your suppliers of AI algorithms and data sets are

Data collection

  • Go beyond the law and regulations and be a prime exmple
  • Directly anonymize any data
  • Enhance data with voluntary submitted data

User interaction

  • Offer users to opt-in or opt-out of data collection
  • Make user actively interact with AI instead of automating everything

Care

  • Be transparent about how data is collected, stored, used and processed and how the AI application works in layman’s terms.
  • Provide evidence for transparency via marketing FAQ page
  • Gather feedback from internal stakeholders and users, re-adjust and improve based on feedback

Marketing stack with AI components

LinkedIn

Marketing purpose

  • Display ads with marketing messages on LinkedIn platform.

AI component purpose

  • Display anch.AI ads to LinkedIn users that are interested in ethical AI.
  • Display anch.AI ads to LinkedIn users with propensity to click on ads and create an account at anch.AI platform. Propensity score is set by a calculation of historic usage of LinkedIn.

Risk assessment

  • Bias risk acceptable due to high data privacy using only user generated data

AdForm

Marketing purpose

  • Display ads with marketing messages for visitors that have previously visited anch.AI (retargeting).
  • Buy ad inventory from publishers.
  • Serve and host ads for display.

AI component purpose

  • Brand protection algorithm to prevent anch.AI ads to display on tragic, controversial and otherwise offensive articles.

Risk assessment

  • Bias risk acceptable because of low and hard to predict impact

Oracle Contextual Targeting

Marketing purpose

  • Display ads with marketing messages on relevant articles at publishers.

AI component purpose

  • Probabilistic match algorithm to match relevant articles with anch.AI ads using word grouping from crawled articles with selected keywords. For example, an article about AI technology might be selected to display an ad for anch.AI.
  • Brand protection algoritm to prevent anch.AI ads to display on tragic, controversial and otherwise offensive articles.

Mailchimp

Marketing purpose

  • Send newsletter email to people registered to anch.AI email list.

AI component purpose

  • None used.

Google Analytics

Marketing purpose

  • Collect aggregated and anonymized user data on anch.AI to create reports on website behavior and traffic sources.

AI component purpose

  • None used.