You surely remember Tom Sawyer. He became famous in economic psychology with his ingenious idea: he reframed the hated chore of whitewashing the fence as an exclusive privilege, convincing his friends to pay him to do his work instead of him.
A recent MIT Technology Review article describes an AI model whose data source is an extremely sensitive area: inmate communication. Furthermore, it explicitly exploits the Tom Sawyer Effect.
However, the ethical dilemmas associated with this are far more general; they do not remain within the walls of prisons.
What risk do you take on as a leader when the algorithm starts analyzing thoughts instead of actions?
Let’s briefly examine two interesting aspects: 👇
1. The Tom Sawyer Effect: Data Exploitation in Close-up 💰
The Original Paradox: Inmates have always paid for outgoing phone calls, which have been recorded and analyzed using certain methods. That’s nothing new. So what has changed?
The system has leveled up: AI has been deployed.
The transcripts of these inmate conversations are now used on a massive scale to train AI models, which then attempt to determine the probability of potentially planned crimes being realized from current conversations.
All costs related to the development and operation of the model are passed back to the inmates.
We understand this is a special situation in prisons.
However, AI-based analysis of communication data is rapidly spreading in other environments as well.
What about workplace communication data?
Serious ethical and legal questions arise concerning the use of workplace data:
Workplace Communication Data Assets: Who owns your employees’ Slack/Teams messages, emails, and even their general communication?
Usage with Consent: You might hold the stance that your colleague does not legally own the email they wrote. But they are certainly the author. How much say should they have in how the communication they generated is used? For example, should their express consent be required to use it for developing AI models that will then be used to monitor them?
Goodwill Risk: What is the risk if ambiguities prevail in your company regarding the above?
2. The Temptation of the Thought Police: Detecting Intention 🔮
The other interesting aspect is this: the “prison AI” was built to “detect and understand when crimes are being thought about or contemplated” – Kevin Elder, president, Securus Technologies.
What about in corporate life?
Crossing the Line: Where is the boundary between early, even “preventive” (?) fraud detection practices and the psychological profiling of employee communication?
False Positive Alerts: A frustrated internal message (“I’ll screw up the whole thing”) might enter the risk matrix, which can affect promotion or even lead to dismissal.
Legal Risk: If dismissal is based on a secret algorithm’s signal, serious legal and compliance liability arises for HR and management. “The technology is way far ahead of the law” – Corene Kendrick, deputy director, ACLU’s National Prison Project.
💡 For Business Leaders: The Currency of Trust
If you want the best talents to choose you and stay with the company, your ethical leadership must be at the center of your values.
Lead with Ethics: Show that you, as a business leader, operate transparently and use AI to support human potential, not for surveillance.
Have an AI Ethical Framework: Integrate an AI Ethical Framework into your HR policy.
Business and HR Leaders:
Where does your company draw the line in collecting employee data?
Do you see the danger when machines start “reading” human intentions in the workplace?
I look forward to your thoughts and to sharing this discussion! 👇
#AIEthics #HRRisk #EmploymentLaw #Recruitment #LeadershipDevelopment
Leave a Reply