Dive Brief:
- Artificial intelligence engineers are cognizant of the ethical risks that the systems they build pose, but they lack the authority, incentives and support needed to adequately address those ethical risks, University of Manchester research says.
- Many AI professionals told Manchester researchers about the safeguards they would like to put in place, but were unable to do so.
- This research was presented at the 10th Data for Policy Conference, which took place from Sept. 8 to 10 in Barcelona, amid the fallout from engineer Jacob Coxon’s viral X thread, in which the engineer loudly resigned from Anthropic and raised his own questions about the morality of AI engineers. HR’s use of AI software, made by these very same engineers, becomes increasingly commonplace.
Dive Insight:
For its study, Manchester researchers interviewed software engineers across sectors, including finance, AI research and semiconductor manufacturing. They summed up the dilemma of AI engineers’ position as having “ethical awareness without ethical agency.” These findings were emblematic of the work experience Coxon described during his stints at Anthropic and OpenAI.
“Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives,” Coxon said in his Sept. 8 post on X. He went further to say that AI systems will soon be able to “hack anything, revolutionize any field overnight, and acquire real power and resources.”
Shortly after this post, Coxon described the exact ethical dilemma Manchester researchers highlighted. “The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt,” Coxon wrote. “If anything, many executives and senior researchers will couch their phrasing in the press to sound sensible – but I hear the same people express fear privately.”
Coxon’s description of the workplace culture at Anthropic and OpenAI mirrored the Manchester research, which should sound some alarms for HR professionals. Among myriad studies that showcase workers’ AI understanding gap and managers’ fraught relationship with AI, a recent report from talent management firm Talogy suggested that AI itself is outpacing human readiness for the technology — even human workers’ ability to understand AI’s utility to future workflows.
So, what can HR do? Prioritize human-ness, even as firms embrace AI, experts recommended, including SHRM leadership. Writing a concrete AI ethics policy can also safeguard against the kind of awareness-without-agency workplace culture that Manchester researchers — and Coxon — described.