A few months ago, I randomly broke it to my editor that some of our chats on the Teams app had a lift from AI. I was not using AI tools to outright automate all aspects of my work. That would be unethical, of course, and a blatant abuse of contract.
I was, instead, using Apple Intelligence to straighten my sentences, fix my typos, and tone down my em-dash zealotry. On two separate occasions, I narrated my message, transcribed the audio with an AI tool, and used OpenAI’s GPT-4 wizardry to proofread the wall of text.
Typing the same message would have taken me at least twice, or thrice, as much time on a phone. If a phone wasn’t handy, I would’ve had to make a platform stop and type the response on my laptop, perched on a buzzing station’s cold metal seat.
As of today, AI has become an integral part of my inbox, and in a way that I feel is more “human.” Before I said “yes” to Gemini becoming a part of my inbox, I was intensely debating whether AI should even exist in a place as intimate as Gmail.
Gemini is there. You just need the leap of faith
“Tell Tyler that the idea is cool, but I don’t cover breakthroughs in drug discovery. Tell him to keep sending pitches, though. Wish him luck.”
This was my most recent prompt to a communication officer at an esteemed science institution, who sent me a press release about a drug that can start working on migraines almost instantaneously.
The email generated by Gemini was an elaborate response that conveyed my excitement for the discovery while expressing my inability to write about a subject that falls outside my area of expertise and the publications’ coverage domain. I felt better after hitting the send button.
It all happens in a few seconds.
Before using Gemini’s compose feature in Gmail, I had rarely ever written emails to reject a pitch, even though I wanted to extend the courtesy of a response and hoped to preserve long-term contact with the sender.
Usually, I try to keep the responses as casual as possible, but if I change my mind, Gemini offers a one-tap route to formalize the words. The “Elaborate” and “Shorten” quick actions also come in handy.
I often throw in a line such as “keep the tone light-hearted” to shoehorn a few words worth a chuckle. The jokes are predictably bad, but they hit the contextual bull’s-eye more often than not.
Otherwise, I am always ready with the backspace, the lifeline for any person who writes for a living. It never hurts to take matters into one’s own hands. With generative AI workflows, that’s more of a necessity than optional caution.
Yes, AI cooks up facts. With Gmail, it’s limited to the email context.
Another reason that I love Gemini is the time it saves me. Being part of a newsroom means things are a little too dynamic. Honestly, it’s a pretty weird spot to be in since AI has had a visibly detrimental impact on the media business.
The rush to cover the latest turn of events has often resulted in ignored emails, either due to time constraints or just because I didn’t have the energy left to handle inbox duties after filing back-to-back stories. That’s where Gemini comes to the rescue.
Now, I’m not a fan of the email summaries, but they do help me decide whether I should proceed to dig deeper into the real content. Once I open an email and choose to respond, Gboard’s dictation feature and Gemini’s smart compose take over the reins.
The whole pipeline is noticeably faster, and at the end of the day, that’s what matters — especially when working across a time zone where 1 a.m. is when the news cycle heats up, and so does the onslaught of emails.
Is it justifiable?
A healthy part of my work routine involves pitching to at least a dozen editors at various reputed outlets every month. Unfortunately, I’m not the only one in the fray. That means the destination inbox is constantly flooded. And here’s the cruel part.
Barely a few pitches get accepted, but most emails land in the no-response black hole. In the early days, I used to think that an esteemed editor should at least have the courtesy to reply with a curt “no,” if not a polite rejection mail. Realistically, that’s not feasible.
Yet, to this day, I feel the sting of a no-response for pitches that I worked hours researching, writing, and fine-tuning. Interestingly, it’s not a hard pill for me to swallow because I find myself in the same spot as the editor. At least a dozen times each day.
My inbox is also a tale of pitch avalanches, ranging from some publicist requesting a guest post from an Armenian crypto evangelist to a company proposing a review for their latest safety smartwatch for kids. I don’t have kids. I don’t understand cryptocurrencies, either. I chuckle at the insanity of the situation and move to the next email.
Yet, I can’t help but feel that the sender deserved a response from me. An approval or a courteous rejection. After all, that’s basic decency, baseline professional conduct, and respect for human dignity.
If, however, I carry that righteous burden and sit down to write an email, I would burn valuable time and energy that would otherwise be well spent on real work.
But if I use an AI agent to do human work, would that retain any meaning? I skirted that dilemma in favor of getting the job done. Now, I use Gemini as an agent to draft responses for me.
To put it more accurately, I need to refine my crude words and shoulder the chore of linguistic accuracy in responding. Yet, the question of ethics still lingers, and depending on the person you ask, using AI for communication can elicit wildly different takes.
I am in a similar dilemma, and to feel easy, I often revisit this passage published in the Journal of Business Ethics. Titled The Ethical Implications of Artificial Intelligence (AI) For Meaningful Work, it’s worth a read:
“When AI assumes simple or complex tasks that workers find boring or repetitive, then this potentially promotes autonomy by freeing up time for workers to build their autonomy competencies through doing other more challenging or authentic work. For example, if an AI prioritizes a worker’s emails so that she only sees those requiring a response, this may free her to work on other more valuable tasks.”
The bigger picture
The sum here is that if my usage of AI hastens the more mundane parts of my job and leaves enough headroom to engage in more meaningful work, it is justifiable. But this is a two-way street, and the person on the other end of the loop may not share those opinions.
Husain Aanis Khan, an Alex Chernov Scholar at Melbourne Law School and an expert in legal design and regulation, tells me that he would prefer short, typo-laden, or even gruff responses instead of long and flowery emails that aren’t a fruit of human efforts and ingenuity.
“What is lacking, is the human touch. That’s what most of us seek, or would prefer,” Khan tells me. When I told him I had trained a writing assistant (aka Gem, as per the Google AI lexicon) on samples of my own writing, resulting in email responses that sounded just like me on any given day, his stance softened.
A paper published in the Journal of Autonomous Intelligence also stresses the importance of the proverbial human touch. It postulates that when AI apps are being widely used, they come at the cost of real human interaction and lose out on the emotional connection.
It’s just AI shaping my intent. Not dictating it.
What if he is never able to guess that AI has a fundamental role in our correspondence? “It is likely. And in that case, I would have no qualms,” Khan said after I showed him how I can adjust the tonality of my emails, making them eerily similar in their jovial tone to the way I joke around with friends.
But here’s the real question. In my self-admittedly ethical quest to speed (and ease) up my workflow, am I unwittingly undermining the person on the other end? Does a person deserve to receive messages that have been given shape by an AI?
“Ignorance is bliss,” I tell myself. Maybe I am choosing the lesser evil between leaving a person on “read” and answering their request. It’s a matter of where your conscience sways. For me, I chose the side where I feel more human in my conduct.
I’ll respond (with some help from AI). It’s not quite the pristine human touch, but the end result still manages to connect humans in a definitive two-way conversation. I’d call that a win at the end of the day.
Read the full article here
