Chamorro-Premuzic gave a fascinating keynote about the ethics of artificial intelligence (AI).
There’s no doubt that people have been concerned about AI causing bias in the workplace and hiring process.
However, Chamorro-Premuzic explained that AI “has the ability to find identify patterns and extremely large sets of data at a very high pace.
“But ultimately, no matter how far we want to stretch dystopian or utopian views of killer cyborgs or weapons coming and taking over the world, or some machine that will improve our work success; AI is still a prediction engine.”
This is where the importance of people comes in and “it’s up to us to determine what we do with that prediction, and how that prediction is incorporated or squeezed in into a business-relevant or word–relevant process.”
On the back of this comment, Chamorro-Premuzic discussed how AI can impact the recruitment process.
Hiring with AI
He noted, “You can implement recruitment processes based on AI, because you want to make your organization more talent-centric, more meritocratic, and ensure that people who are better or seem better suited for a job get hired or promoted internally.”
Nonetheless, “an unintended, unethical consequence of that might be that you make things even harder or worse for those who are less privileged.”
With this in mind, AI still needs to be monitored and regulated. But Chamorro-Premuzic had some interesting thoughts on acknowledging and overcoming machine bias.
Bias in AI
Chamorro-Premuzic contextualized bias in AI: “When companies, whether it’s Microsoft or Amazon, decided to train a chatbot to predict who gets promoted in a certain team or environment, that AI engine learns very quickly, that based on the prediction, the candidates with the highest potential to succeed in an organization are more likely middle age, male engineers.”
He noted that this issue is spawned from those who code the AI, and that the machine cannot develop this kind of bias itself: “AI tries to learn how to act in a human way, it will also reproduce the bad, or the dark side of human thinking and human decision making.”
When it comes to addressing this situation, Chamorro-Premuzic recommends looking at the teams who create and train the AI.
In terms of overcoming or acknowledging bias while improving process, Chamorro-Premuzic was positive.
“The good news is that we have an opportunity to learn from this and use algorithms and AI in a way that actually de-bias existing decisions and existing processes,” stated Chamorro-Premuzic.
Tomas Chamorro-Premuzic’s whole discussion is undoubtedly a much watch for anyone interested in the role of technology and AI at work.