Letter: Humans and machines should both up their game

I loved Sarah O’Connor’s trenchant column “AI is making making use of for work much more depressing” (Opinion, November 30).We accurately measure AI in opposition to an ordinary to which we may by no means maintain a human. Human interviewers will be — your readers could decide how steadily — impolite, lazy and unsympathetic, topic to bias both acutely aware and unconscious. So blended is the observe report that one of many strongest criticisms of AI instruments is that they bake in historic human efficiency bias. The alternative that AI methods supply is to behave as a mirror to the selections hiring organisations take. If set up and managed accurately, they measure what occurs and can quantify what truly does and doesn’t work for numerous mixes of candidates.The key to getting this proper is evident transparency round knowledge, conserving people within the loop and a deal with explaining AI methods to stakeholders. Research means that we’re extra keen to accommodate poor efficiency from a human than a machine. Perhaps — however higher certainly to cross-fertilise to enhance both.Tim GordonPartner, Best Practice AILondon N1, UK