BYOAI? How to bring your own AI to work
How to use AI (responsibly) and protect yourself from burn-out
In the 2024 Work Trend Index published this week by Microsoft and LinkedIn, the company’s survey of 31,000 workers across 31 countries found that 75% use AI and 90% of users say it makes them more efficient. Perhaps even more interestingly, 71% say they’d hire a less experienced candidate with AI skills over a more experienced candidate without them.
For now, let’s ignore the fact that I don’t think “AI skills” are just about learning to use ChatGPT or similar tools.
Many people have asked me how to effectively use AI at work to help avoid burnout or even take better care of themselves once they have recovered.
Here is a concise summary of all that I learned from balancing a PhD, multiple freelancing projects, writing a book (The Ethics of AI), and parenting during the pandemic – the successes, as well as the failures, of using AI to prioritize purpose, progress, and passion over hours logged.
Some life news
After six years in Abu Dhabi, in the UAE (a great country I recommend everyone to consider relocating to, BTW), I’m excited for my family's next adventure in Sydney. The past couple of weeks have been a whirlwind of goodbyes, lots of tears, packing up our home, and trying to savour our final days in a city that will always feel like a second home. Before that, while my wife was already in Sydney finding our new home, I spent a wonderful month traveling through Italy with our kids. While the transition to a new country is never easy, we're looking forward to embracing all that Sydney has to offer.
Why am I telling you this?
The latest advancements in AI and its fast (too fast, though) penetration into any software have been invaluable tools in making this period feasible with my commitments. From scheduling meetings across multiple timezones to writing emails and preparing for many talks I gave in Italy, AI assistants have streamlined many of the tedious tasks involved in having to work while organizing an international move.
I’m busy-ish, and more and more, I’m using AI tools to support me. My use of AI has evolved significantly in just a few years. During the pandemic, I juggled my PhD, freelancing, writing, coding, and parenting. This was before the LLMs revolution. Back then, it was all about unwavering focus on time management and acceptance of the 70-20-10 rule: 70% of your work will be mediocre, 20% will suck, and 10% will be amazing. Just accept it. It will liberate you! (Thank you, Ben Meer and Jonathan Mann for the discovery)
What began as using AI writing tools for first drafts and outlines has expanded into leveraging AI for writing software, custom agents for my tasks, project management, data analysis, and even creative tasks like generating ideas and brainstorming. AI has been a force multiplier for my productivity while allowing me to focus on the higher-level aspects of my work that require human judgment, creativity, and compassion.
So, how do you bring your own AI to work?
AI has gone mainstream, with tools like Claude, GPT-4o, and Google's latest offerings presented at I/O rapidly expanding what's possible with AI assistance at work. Used effectively, AI can be a huge productivity multiplier - I've calculated saving over 10 hours per week by leveraging AI for coding, writing, research, and project management tasks.
But how can you actually start incorporating AI into your own workflows? Here are some practical tips:
Identify repetitive tasks that are bogging you down. Things like writing emails, data entry, documentation, meeting notes, report writing, and scheduling are all prime candidates for AI assistance. Make a list of your most tedious, time-consuming processes.
Evaluate each task's suitability for AI tools:
How structured and well-defined are the inputs and outputs?
Do you have sufficient data/examples to "train" the AI?
Is there any sensitive data that could raise privacy/security concerns?
For suitable tasks, pick a specific AI tool and run a pilot test:
For writing: Claude, ChatGPT (even the free version, 3.5 is still great, but the writing style of Claude is much better)
For coding: GitHub Copilot, Code Llama
For data/analysis: ChatGPT-4, now 4o is even better
For research: Perplexity
There are thousands of transcribers for meeting notes. I personally like MeetGeek, though they are not great yet. Often, their summaries miss the point that a human would rather save or underline. But it saves time writing, redacting, and cross-checking notes.
For email: Superhuman! I usually give Superhuman AI bullet points using my keyboard voice transcription features, and email writing goes from 5-10 minutes per email to 1-2 minutes. Compound that for 10-20 emails per day!!
ALWAYS closely evaluate the AI's outputs—don't just blindly trust them. Look for errors, bias, nonsensical outputs, inaccurate citations or facts, misunderstood emotions, and what emotion you want to convey to your audience.
Iterate by feeding more examples to the AI, fine-tuning prompts, and monitoring performance. Anthropic just released a fantastic tool: give it a badly written or dictated prompt, and it will give you the proper prompt you have to use.
Once comfortable, integrate the AI into your regular process, but with a human-in-the-loop to review all outputs.
Finally, let me recommend disclosing when content is AI-generated; never pass it off as your own work. Carefully vet tools for security and restrict access to sensitive data. We're in uncharted territory, so document everything for auditing and get stakeholder buy-in.
AI opens up amazing possibilities for streamlining work, but we need to intentionally design human-centric processes that prioritize fulfilling work over pure productivity. Don't just add AI to existing bloated processes and long hours. Use the efficiency gains to reinvest time into more strategic, creative, and collaborative endeavors that leverage our unique human strengths.
This is just the start of AI's impact on work. We can rethink jobs by bundling "human" skills like emotional intelligence, problem-solving, and ethics with AI's analytical prowess. By responsibly bringing AI to work now, we can help shape that future of work.
Creating the AI playbook
For organizational leaders, the challenge will be cultivating a culture that embraces continuous learning alongside AI adoption. How can we focus on developing skills-first playbooks that prioritize relevant skill acquisition over traditional role requirements?
Crucially, how do we ensure that new job seekers stay engaged and gain real-world wisdom?
My children are 11, 7, and 2.5, but if they were graduating this year, I would advise them this.
Invest in your learning: Double down on your curiosity to learn about the unique challenges and opportunities at your organization.
STEM isn’t everything: Adaptability will be key, and developing the skills for empathy, critical thinking, common sense, judgment, and leadership will always serve as an anchor for the new job market.
Impactful use of AI: Finally, it’s crucial for new entrants to the workforce to consider how they can use AI not just for personal career advancement but to ask, what change do you want to make in the world? Whether it’s addressing social justice or enhancing public health, AI offers tools that, when used ethically, can significantly contribute to global wellbeing.
This is what I think it’ll be like for life, I think – understanding how we integrate new technologies into our day-to-day jobs, and use the time saved to explore how we bring additional value or allow ourselves to maintain productivity while reducing our digital debt.
Thank you so much Alberto and good luck to you!
Benvenuto a Sydney!