Mike FrenetteManager, IT PMO| Halifax Water (retired)Halifax, Nova Scotia, Canada
Many generative AI products will take your input and apply it to what it finds in its data lake, then generate something that will save you time and effort. But will it be accurate? Will it be unbiased? Will it be the same as you would have written? How do you feel about that? Is it ethical? If so, how do you make it our own? Saving Changes...
Anjana KakuAssociate Director - Program Management - Data & AI Services| Kyndryl Solutions Pvt. Ltd.Bangalore, Karnataka, India
Mar 09, 2024 5:34 AM
Replying to Sergio Luis Conte
...
What generative AI does is to automate something that you of others could do by hand. So, in this case, is a matter of process automation is you like to see it in this way. For example, you can do something similar using no-code or low-code tools. On the other hand, usually people forget that the results has a confidence score associated to it. It does mean, it has an inherent error so the "human in the loop" concept must be applied when using generative AI.
Even the low code no code tools use AI in the background. So whatever system is at work, human-in-the-loop is important. We can depend on the tools to make our lives easier but the accountability of the output / outcomes of the tools would be ours. Even the confidence scores are calculated based on certain parameters/data that could have an inherent bias, and hence the confidence scores may not be accurate. Complete dependence on AI systems (low code no code included) is not the way to go.
...
1 reply by Sergio Luis Conte
May 18, 2024 7:17 AM
Sergio Luis Conte
...
You right. That´s my point. And it has been proved from more than 40 years ago.
Saving Changes...
Anjana KakuAssociate Director - Program Management - Data & AI Services| Kyndryl Solutions Pvt. Ltd.Bangalore, Karnataka, India
Mar 10, 2024 12:46 PM
Replying to Aung Sint
...
Mike, I agree with the other contributors in this discussion. We use tools and techniques to improve our efficiency and effectiveness, so there's nothing wrong with leveraging the tools that are available to us. It would be unethical to claim that something isn't our own just because we used a tool to complete it. It's our responsibility to verify the accuracy of the output and provide a reference to the source of information. In my opinion, humans tend to be biased in general (though this statement itself could be biased!).
Humans tend to be biased, and so are the AI systems, but what makes humans better is the fact that they understand that they could be biased, but AI systems / models do not, and that intelligence is what sets humans apart. So yes, we could use tools to ease our work and that would not be unethical, but we'd have to take accountability of the outcomes, and so reviewing the output of the tools is our responsibility. And then we retain the right to claim that the contribution is ours. Saving Changes...
Anjana KakuAssociate Director - Program Management - Data & AI Services| Kyndryl Solutions Pvt. Ltd.Bangalore, Karnataka, India
Mar 08, 2024 12:50 PM
Replying to George Freeman
...
In followup to my statement, here are the specific ethical concerns:
PMI Code of Ethics:
- Honesty: Mandatory Standards, Section 5.3.1: We do not… state half-truths or provide information out of context.
- Responsibility: Aspirational Standards, Section 2.2.4: When we make errors or omissions, we take ownership and make corrections promptly.
The opportunity to violate this standard dramatically increases when a project manager uses tooling to generate content. In some cases, it might take longer to validate/confirm generated content than creating it from scratch yourself. Context is everything in project management; as we should know, this is the “kryptonite” of Generative AI.
Do you agree that these PMI Code of Ethics sections are relevant to this discussion?
George, using a tool to ease our work might not be unethical. I might create a project document using a Macro or AI, it is upto me. What matters is the project document & its accuracy. Information about the tool used to create project documentation might be out of context (this could be debatable though). However, the responsibility for errors / omissions in the output document is entirely ours. And once those have creeped in, we're required to take ownership and make corrections promptly. That according to me is in line wiht the PMI code of Ethics: Honesty & Responsibility. Saving Changes...
Anjana KakuAssociate Director - Program Management - Data & AI Services| Kyndryl Solutions Pvt. Ltd.Bangalore, Karnataka, India
Mar 07, 2024 9:07 PM
Replying to Akash Bhatt
...
Well, Mike, I don't think it is ethical since the AI is doing your work, but as we have it, it is better to generate ideas at least and then modify them according to your needs when you are a beginner; once you get experience in making documents, you can become far better than AI since it will not be biased and not monotonous document. It will be professional and unique to that. situation
It isn't a question of ethics when a PM uses AI to generate a document or do some other work. What is important is reviewing the outcome and owning it, irrespective of whether the outcome is positive or negative. Lets look at it this way - A PM might use a macro to do some financials, instead of manual calculations, and that would not be unethical even if the PM has not declared the use of macros, and similarly if the tool in question is AI, it does not become unethical. It still is the PM doing the work, provided the PM stays accountable. Saving Changes...
Anjana KakuAssociate Director - Program Management - Data & AI Services| Kyndryl Solutions Pvt. Ltd.Bangalore, Karnataka, India
I'll say, the Gen AI models are not perfect. Few would be better than few others. It is upto the producers of the Gen AI models to declare the confidence scores for the models they create, so PMs can decide what to use when needed. It does save time & efforts, but the output of the models cannot be claimed to be accurate or unbiased as the output depends on the training data provided to the models which would not be accurate or free from bias. It might also be different than what we would write in most cases. The output could be surprisingly better than my expectations or could be totally useless. But so long as I do not blindly trust the output and use it as is, I am totally in control. I can pick that output, enhance it, modify it to what I understand is correct, add my own thoughts and content to it, and then share it with the others. This is how I make it my own. I totally own the output now, and I am accountable for it. I am still following the PMI Code of Ethics, and am in alliegence with it.
...
1 reply by George Freeman
Mar 20, 2024 10:51 AM
George Freeman
...
Anjana,
With the monetization, extraordinary investments, and societal focus on AI-based technologies and services, we will see accelerative degrees of change that may outpace our ethical constructs in society, let alone those housed in our PMI Code of Ethics.
Consider this progression:
- AI as a tool (current state) >> AI as an assistant >> AI as a proxy >> AI as a ?
In our current state, with AI (e.g., GenAI) leveraged as a tool, our risk of ethical exposure is reasonably mitigatable through commonsense content accountability (as you have portrayed). However, that does not mean we shouldn’t highlight the (current day) risk potential in this regard.
My more significant concern is the above progression. Will our ethical constructs have the same foundational value when “AI as an assistant” becomes the standard presentation of AI? What will happen as we move towards “AI as a proxy” and start seeing the blurred lines between virtual and physical representations?
The past is our lens to the future societal re-baselining of norms. If we do not challenge the life-cycle perspectives and ethical considerations now, we will find ourselves without a compass in the future.
As other contributors have brought up, using AI tools can make people work more efficiently but AI can also give you different points of view on the topic, and get you started faster. AI creates very good content but AI doesn't understand the context, environment, company, and people for example related to the project. The content created by AI should be always reviewed and modified based on the situation. This is the place where the project manager creates value.
One thing I've found helpful is providing GenAI with context, asking it to review the content and then respond to my prompt. Technically, you could give it context, environment, company, and people (roles, responsibilities attitudes, interests, authority...) related to the project. Too much detail about your company or people involved could raise ethical or privacy concerns, and you'd still want to review it to make sure GenAI doesn't get too creative (because it can). Saving Changes...
George FreemanThought Leader | Author | Architect| Florida, United States
Mar 20, 2024 8:40 AM
Replying to Anjana Kaku
...
I'll say, the Gen AI models are not perfect. Few would be better than few others. It is upto the producers of the Gen AI models to declare the confidence scores for the models they create, so PMs can decide what to use when needed. It does save time & efforts, but the output of the models cannot be claimed to be accurate or unbiased as the output depends on the training data provided to the models which would not be accurate or free from bias. It might also be different than what we would write in most cases. The output could be surprisingly better than my expectations or could be totally useless. But so long as I do not blindly trust the output and use it as is, I am totally in control. I can pick that output, enhance it, modify it to what I understand is correct, add my own thoughts and content to it, and then share it with the others. This is how I make it my own. I totally own the output now, and I am accountable for it. I am still following the PMI Code of Ethics, and am in alliegence with it.
Anjana,
With the monetization, extraordinary investments, and societal focus on AI-based technologies and services, we will see accelerative degrees of change that may outpace our ethical constructs in society, let alone those housed in our PMI Code of Ethics.
Consider this progression:
- AI as a tool (current state) >> AI as an assistant >> AI as a proxy >> AI as a ?
In our current state, with AI (e.g., GenAI) leveraged as a tool, our risk of ethical exposure is reasonably mitigatable through commonsense content accountability (as you have portrayed). However, that does not mean we shouldn’t highlight the (current day) risk potential in this regard.
My more significant concern is the above progression. Will our ethical constructs have the same foundational value when “AI as an assistant” becomes the standard presentation of AI? What will happen as we move towards “AI as a proxy” and start seeing the blurred lines between virtual and physical representations?
The past is our lens to the future societal re-baselining of norms. If we do not challenge the life-cycle perspectives and ethical considerations now, we will find ourselves without a compass in the future.
George Saving Changes...
Ming YeungAdjunct Professor & Acting COO/CPO/CRO (contract)| Blockchain Venture Capital Inc.Toronto, Ontario, Canada
Hello Mike, what a timely topic to deliberate.
Generative AI products indeed offer a fascinating blend of convenience and complexity. Users should be mindful of these topics which can exhibit these attributes worthy of consideration.
-- Accuracy: Generative AI models learn from vast amounts of data, but their accuracy depends on the quality and diversity of that data. They can be remarkably precise, especially for repetitive tasks or pattern recognition.
-- Unbiased Output: However, biases present in the training data can seep into the generated content. For instance, if the training data contains gender or racial biases, the AI might inadvertently produce biased results.
-- Originality: While generative AI can mimic patterns, it lacks true creativity. It doesn’t invent new ideas; it remixes existing ones. As humans, our unique perspectives and creativity remain unmatched.
-- Personal Touch: When you write something, it carries your voice, style, and context. Generative AI can’t replicate that personal touch. It’s like comparing a mass-produced painting to an artist’s original masterpiece.
-- Ethical Bias Mitigation: To address bias, we must curate diverse training data and actively debias the model. Regular audits and adjustments are crucial.
-- Human Editing: Use AI-generated drafts as a starting point. Inject your voice, add context, and refine it. The blend of AI assistance and human touch can yield powerful results.
In summary, while it is powerful, generative AI is a not a replacement for human creativity, empathy, and originality. As creators, we shape AI, and in turn, it shapes us. Let’s wield it ethically and responsibly.
Saving Changes...
Ryan LamingRegional Director of System Integration| LumeoOntario, Canada
Using GenAI to produce project documents which the human/PM then reviews and refines is a time-saving and productivity-enhancing method. AI is a fact of life at this stage, and not using it is an impediment to driving organizational and project success.
The ethical piece to this is ensuring proper resources receive attribution, as applicable, but using AI for the purposes of this specific thread/question is entirely ethical. Saving Changes...
Khanh NguyenProject Manager| Parcel PerformHo Chi Minh City, Sg, Viet Nam
I think there's nothing wrong with using AI to support your work as long as you don't expose sensitive data to them. Saving Changes...
Information is not knowledge,
Knowledge is not wisdom,
Wisdom is not truth,
Truth is not beauty,
Beauty is not love,
Love is not music
and Music is THE BEST
Community Champion
