ICO: Gui­de­li­nes on AI and data protection

The Eng­lish regu­la­to­ry aut­ho­ri­ty ICO has Gui­de­li­nes on AI and data pro­tec­tion published:

  • Part one addres­ses accoun­ta­bi­li­ty and gover­nan­ce in AI, inclu­ding data pri­va­cy impact assess­ments (DPI­As);
  • Part two con­cerns fair, law­ful, and trans­pa­rent pro­ces­sing, inclu­ding legal bases, eva­lua­ting and impro­ving the per­for­mance of AI systems, and miti­ga­ting poten­ti­al discrimination;
  • Part three con­cerns with data mini­miz­a­ti­on and data secu­ri­ty; and
  • Part four con­cerns the gua­ran­tee of data sub­jects’ rights, inclu­ding in the con­text of auto­ma­ted decision-making.

Cf. also the ISO stan­dard for trust­wort­hi­ness of AI systems.