We’re witnessing an increase in product articles the place individuals are overtly proud about saying that they wrote their article with ChatGPT. Is that this the way forward for product administration journalism?
Tright here isn’t any method round it — product managers should be nice at writing. It’s a talent that may not be faked or handed off to another person. I perceive, as a product supervisor you’ve a number of priorities and never sufficient time to do them, however relating to duties you can delegate — writing is just not one among them.
You the product supervisor are anticipated to be proficient at organizing your considering, your studying processes, experiments, your knowledge, and most significantly, your classes discovered and insights; And there’s no higher demonstration of that than you producing writing that’s clear and arranged, and achieves goals for all who learn your materials.
So what does it say about you as a product supervisor for those who determine to let AI do your writing for you?
I’m exploring this query as a result of already, we’re witnessing an increase in product administration articles and different types of writing from people who find themselves overtly proud about saying that they wrote their article with ChatGPT.
Now why these “writers” consider it’s a good suggestion to have ChatGPT write an article, after which they publish it and take credit score because the “creator” — I’m admittedly inquisitive about their causes. However to see this pattern emerge on the planet of product administration journalism does give others and myself causes to be involved about the way forward for the product administration craft.
Hey, we get it. Writing is difficult. Generally it’s arduous to start an article and also you simply want that one concept or perhaps just a few opening strains to get your inventive juices flowing.
However what occurs when the necessity for just a few opening strains remodel right into a full first draft, then you definitely add just a few strains of your individual, and then you definitely add your title to it because the creator? Do you actually wish to reside in a world the place taking credit score for an AI-written article is the brand new journalism?
This isn’t nearly the way forward for writing. It’s about the way forward for studying, which is the bedrock of product administration.
Our core worth as product managers is to generate and articulate insights from analysis and experimentation. How are we supposed to do this if the majority of the educational by way of writing is just not finished by you, however by AI?
Once you use ChatGPT or any AI expertise to do your writing for you, let’s be clear about what you’re really asking it to do. You’re instructing AI to carry out the investigation and evaluation required with a view to generate, perceive, and talk insights.
We’re not speaking about crunching numbers or analyzing quantitative knowledge units. You’re asking AI to investigate knowledge, perceive context, generate and check hypotheses, discover patterns and proof, make inferences, and based mostly on the outcomes, create tales with a view to make the information imply one thing to folks.
You will have to bear in mind — regardless of how subtle ChatGPT is, it should all the time lack the flexibility to perceive qualitative knowledge, similar to dialog, feelings, creativeness, and different types of solely human types of communication. You may instruct ChatGPT to jot down you a imaginative and prescient to your subsequent product concept, however don’t count on something that’s groundbreaking. All it may do is write a imaginative and prescient for what has already been finished, as a result of the expertise is pushed by knowledge that’s based mostly on what has already been created.
Do you keep in mind the late Ada Lovelace? She is thought to be the world’s first pc programmer, and within the 1800s, she was credited for writing the world’s first machine algorithm. Lovelace was additionally a colleague of the late Cambridge professor, Charles Babbage, the inventor of the Distinction Engine and the Analytical Engine. When Babbage gave a lecture concerning the Analytical Engine, Lovelace was commissioned to translate into English, the lecture notes transcribed by the late French scholar and former Italian prime minister, Luigi Federico Menabrea.
What’s fascinating about these notes is, throughout that point, Babbage’s friends explored the query of whether or not his Analytical Engine might be thought to be a “considering machine” i.e. an “Synthetic Intelligence” of its time. Intriguingly sufficient, Lovelace didn’t agree with this chance. Listed below are some excerpts from the translated notes:
“Subsequently, since, from the second that the character of the calculation to be executed or of the issue to be resolved have been indicated to it, the machine is, by its personal intrinsic energy, of itself to undergo all of the intermediate operations which result in the proposed outcome, it should exclude all strategies of trial and guess-work, and might solely admit the direct processes of calculation.”
Translation: For AI to do what a product supervisor does and supply insights based mostly on producing hypotheses and conducting experiments, it should expertise the journey of trial and error with a view to study and generate unique insights. Since AI is intrinsically unable to do that, the product supervisor as a substitute goes by the trials and errors. Any quantitative knowledge generated from outcomes may be given to AI to carry out calculations — which isn’t the identical as producing insights.
“It’s fascinating to protect towards the opportunity of exaggerated concepts which may come up as to the powers of the Analytical Engine. In contemplating any new topic, there may be ceaselessly a bent, first, to overrate what we discover to be already fascinating or exceptional; and, secondly, by a kind of pure response, to undervalue the true state of the case, once we do uncover that our notions have surpassed those who have been actually tenable.”
“The Analytical Engine has no pretensions no matter to originate something. It could do no matter we all know methods to order it to carry out. It could observe evaluation; but it surely has no energy of anticipating any analytical relations or truths. Its province is to help us in making accessible what we’re already acquainted with.”
Translation: It’s comprehensible that product individuals are excited concerning the capabilities of ChatGPT, however let’s be real looking. AI is intrinsically incapable of constructing selections based mostly on feelings and unique ideas. Except we as a worldwide society, conform to redefine the which means of “unique ideas”, for AI it’s categorically unattainable. AI algorithms are designed to make sure that selections made are based mostly on predefined guidelines. It could’t innately select to make exceptions to guidelines like a human may.
Predefined guidelines contradict with one of many functions of product managers, which is to find distinctive methods to appeal to customers to services and products. For product managers to do that, they will need to have empathy and perceive the feelings of customers, perceive what they need, after which perceive the “why” with a view to uncover what customers actually want — and therein lies the battle.
AI can offer you data because of computations carried out in your knowledge units, however that’s not what makes the data insightful. Data can solely let you know what occurred, however insights are designed by product managers based mostly on their understanding of the complexities behind the data with a view to clarify why it occurred.
AI would possibly be capable of imitate or give the impression of “understanding”, however in actuality, it can’t, as a result of it should by no means be capable of understand, comprehend, conceive, think about, consider, empathize or hypothesize in methods people can.
That’s the reason, whatever the outcomes you’re anticipating to get from ChatGPT while you instruct it to jot down your article — it should by no means be insightful.
In actuality, you’re making an attempt to coerce AI — if that’s even potential — to undergo the rigors of studying — which for AI can be a simulation. You additionally wish to coerce AI to study about one thing that you simply as a human are anticipated to study for your self. It’s you who must be the knowledgeable in making the connections, ascertaining the which means, and discovering out the “why” inside knowledge so you’ll be able to inform stakeholders the correct tales behind the information.
Product managers are required to be grasp storytellers, but you might be telling the Synthetic Intelligence (AI) to jot down the story for you.
In different phrases — you might be telling AI to inform you the “why”.
Certainly you see by now the counterintuitive nature of utilizing AI to do the very job that nobody else is anticipated to do however you, the product supervisor.
Should you as a product supervisor are going to contemplate buying and selling your capability to study and uncover the “why” — for the flexibility to take credit score for what an AI writes — product administration may not be the correct profession for you.
I’m saying — give it some actual thought. The way forward for product administration and actual studying by human beings rely on it.
Be happy to share your ideas within the feedback, and maybe you shouldn’t let ChatGPT try this for you as effectively.
