Ian WhittinghamManaging Director| Calixo ConsultingGolden Cross, East Sussex, United Kingdom
What are the most important ethical issues facing a project manager when applying AI to their projects? What steps would you take to ensure that AI is applied ethically to your projects? Saving Changes...
Do we need to take steps "to ensure that AI is applied ethically to [our] projects?"
What are you doing with AI that is so different from what you've been doing without AI that you need additional steps to be ethical? What does it mean to "apply AI ethically" to a project?
I would argue that, for many project managers, we don't need to do anything different than we're already doing. The ethics of using AI fits in with existing policies. I don't have actual numbers, but given how adoption usually goes, there are still plenty of PMs not using AI on their projects, and most are probably using it as a glorified Google search. They may be using it to create project artifacts, like charters, instead of downloading a template. There may be security concerns with submitting project information to an external AI, but is that really an ethical concern?
Now, as you start getting into the realms of using automation and autonomous AI as tools to help produce project deliverables, I can see the potential for needing additional steps with regards to ethics, but I think the greatest area of ethical concern is when a project is producing autonomous and automated AI deliverables - a very small subset of project managers, albeit with the potential for huge impact.
Do you see the need for additional steps across ALL projects and project managers? If not, what criteria merits the need for additional steps?
...
1 reply by Ian Whittingham
Dec 06, 2023 8:27 AM
Ian Whittingham
...
Thank you, Aaron, you make a very important distinction between projects that make use of AI as an aid to performing and managing project activities and those in which the project deliverables are themselves autonomous and automated AI artifacts. In determining if additional steps may be required, it always helps to examine the risks involved in the project.
Although their shortcomings soon become apparent when you start applying them to real life, Isaac Asimov’s three Laws of Robotics are always a good place to start when thinking about the ethical implications of what might go wrong when autonomous and automated AI interacts with human beings. (You can find a good critique of these shortcomings in this paper by philosopher Chris Stokes https://www.ijrei.com/controller/client/Wh...not%20work.pdf)
The first law states that “a robot may not injure a human being or, through inaction, allow a human being to come to harm.” Despite the flaws that quickly emerge when this law makes contact with the messy complexities of reality, its most important component is its explicit prohibition on causing harm or injury to human beings.
Before taking the PMP exam, back in 2005, I attended a classroom exam prep course. Among my fellow students were a couple of project managers from a well-known North American pharmaceutical company. During a coffee break, I was chatting with them and remarked that the depth and breadth of their understanding of Project Risk Management--which we had just covered in the previous session--was far ahead of anyone else’s in the classroom. “If your software goes wrong,” replied one of the project managers, “you ruin your customer’s day. But if we get it wrong, we ruin their life. Or worse.”
The steps that we take in our projects to manage risk always need to be calibrated against the degree of potential harm that might result from those risks. For many (most?) of us, the resultant harm will fall well withing the realm of negligible. But for any projects that develop or utilize autonomous and automated AI the potential risk of causing harm or injury is much greater, as we have already seen from some recent incidents involving self-driving cars https://www.theguardian.com/technology/202...driving-cars-gm
Saving Changes...
Ian WhittinghamManaging Director| Calixo ConsultingGolden Cross, East Sussex, United Kingdom
Dec 04, 2023 11:19 AM
Replying to Aaron Porter
...
Do we need to take steps "to ensure that AI is applied ethically to [our] projects?"
What are you doing with AI that is so different from what you've been doing without AI that you need additional steps to be ethical? What does it mean to "apply AI ethically" to a project?
I would argue that, for many project managers, we don't need to do anything different than we're already doing. The ethics of using AI fits in with existing policies. I don't have actual numbers, but given how adoption usually goes, there are still plenty of PMs not using AI on their projects, and most are probably using it as a glorified Google search. They may be using it to create project artifacts, like charters, instead of downloading a template. There may be security concerns with submitting project information to an external AI, but is that really an ethical concern?
Now, as you start getting into the realms of using automation and autonomous AI as tools to help produce project deliverables, I can see the potential for needing additional steps with regards to ethics, but I think the greatest area of ethical concern is when a project is producing autonomous and automated AI deliverables - a very small subset of project managers, albeit with the potential for huge impact.
Do you see the need for additional steps across ALL projects and project managers? If not, what criteria merits the need for additional steps?
Thank you, Aaron, you make a very important distinction between projects that make use of AI as an aid to performing and managing project activities and those in which the project deliverables are themselves autonomous and automated AI artifacts. In determining if additional steps may be required, it always helps to examine the risks involved in the project.
Although their shortcomings soon become apparent when you start applying them to real life, Isaac Asimov’s three Laws of Robotics are always a good place to start when thinking about the ethical implications of what might go wrong when autonomous and automated AI interacts with human beings. (You can find a good critique of these shortcomings in this paper by philosopher Chris Stokes https://www.ijrei.com/controller/client/Wh...not%20work.pdf)
The first law states that “a robot may not injure a human being or, through inaction, allow a human being to come to harm.” Despite the flaws that quickly emerge when this law makes contact with the messy complexities of reality, its most important component is its explicit prohibition on causing harm or injury to human beings.
Before taking the PMP exam, back in 2005, I attended a classroom exam prep course. Among my fellow students were a couple of project managers from a well-known North American pharmaceutical company. During a coffee break, I was chatting with them and remarked that the depth and breadth of their understanding of Project Risk Management--which we had just covered in the previous session--was far ahead of anyone else’s in the classroom. “If your software goes wrong,” replied one of the project managers, “you ruin your customer’s day. But if we get it wrong, we ruin their life. Or worse.”
The steps that we take in our projects to manage risk always need to be calibrated against the degree of potential harm that might result from those risks. For many (most?) of us, the resultant harm will fall well withing the realm of negligible. But for any projects that develop or utilize autonomous and automated AI the potential risk of causing harm or injury is much greater, as we have already seen from some recent incidents involving self-driving cars https://www.theguardian.com/technology/202...driving-cars-gm Saving Changes...
Community Champion
