We talk a lot about AI (artificial intelligence) — I mean a LOT! The Kitchen has seen 58 posts on the topic this year alone, that’s one post or more each week!

Understandably, much of this has focused on the use of AI in scholarly communications and publishing, or its role in our global information economy as a whole. Whether generative or agentic, driven by LLMs (large language models) or not, AI features are popping up in every one of our usual sites and tools — from email and word processing to healthcare and online shopping.

But, are we experiencing efficiency gains first-hand, or seeing any problems solved (big or small)? Have a favorite AI-powered tool you’d recommend? Or are you switching off all the AI applications as they roll out?

Today, we ask the Chefs: What’s your favorite AI hack? Or do you lack a hack and wish AI would stay out of your life?

Lettie Conrad

Generally, I’ve been open to experimenting with AI, but only where I can justify the personal or social expense (e.g., GAI images for everything is a waste of human and environmental resources). I consider myself a fast follower — not the first to try a new gadget or widget, but I’m usually in the second or third wave of new-tech adopters. And I’m fascinated by how people manage their individual realms of information, in scholarly work and beyond.

So far, I’m loving what CoPilot can do to supercharge my use of Microsoft Excel. Asking for a formula I can cut/paste or a prompt to double-check my logic has been saving me hours of trying to master macros and pivot tables all on my own (and the resulting headaches!). It may be rubbish at image generation, but I recommend looking to CoPilot if you’re not a spreadsheet pro.

On the other hand, don’t bother trying to get spreadsheet or data analysis help from Gemini — but it has really saved me time in getting a first draft of documents for product development. I can point it to a project charter, user stories, and a user research report to get a solid start on a feedback form for beta testers.

I’ve found that Claude has saved time for my UX colleagues during ideation for new designs and quick feedback from stakeholders. The more complicated the wireframes, however, the more glitches arise and eat up that time savings with troubleshooting.

Roohi Ghosh

I’ll admit it — I’ve started taking AI for granted. And honestly, I mean that in the best way possible.

When I first (very grudgingly) started using tools like ChatGPT, it was just to see what the big fuss was about! But as I used it, it made my life clearer. My role involves a lot of ideation, and while the ideas are there, sometimes putting those ideas into some sort of structure can become a challenge. You know that feeling when you’re circling around a great insight but can’t quite pin it down or explain it? That’s where AI steps in as a surprisingly patient thinking partner.

There have been times when I was struggling to position something — a marketing campaign, a key message, a framework — and after several back-and-forths with ChatGPT, the fog started to lift. It didn’t “solve” the problem for me; it helped me see it better. It helped me problem-solve, one chat at a time. Sometimes it takes talking something out (even if it’s just with a bot) to find your way through.

I also use AI when I’m working on strategy documents. The ideas, insights, and direction are mine, but shaping them into a story that convinces others? That’s where ChatGPT helps. It gives my thinking a skeleton — a structure that makes it easier for others to follow, and for me to position the strategy with more confidence.

Another area where this is a blessing is reporting impact. Whether it’s a monthly, weekly, or quarterly report, there are times when the same information needs to be presented in multiple ways: a detailed quarterly report, a short deck for leadership, a summary for our intranet. AI helps me adapt tone, style, and format without losing the core message. Such a time saver!

And then there’s Paperpal — CACTUS’ own AI tool — which I use for the final polish. It’s an extra layer of assurance. All Editage Insights articles also go through Paperpal for grammar,  consistency, and plagiarism checks, like an AI “reviewer” that helps polish the work, not change it. Since I’m part of CACTUS, I acknowledge this could be seen as a partial choice, but it’s genuinely a strong tool that I’d recommend for its reliability and precision.

I know some people hesitate to use AI, but here’s how I see it: the thoughts and ideas are mine. The originality is mine. If a tool helps me express them better or faster, why wouldn’t I use it?

We don’t feel guilty about using calculators (unless we’re mathematicians, maybe). We use vacuum cleaners instead of brooms, blenders instead of mortars and pestles, and Excel formulas instead of manual calculations. These tools make us more efficient — not less capable.

So yes, I’ve started to take AI for granted. I reach for it naturally when I’m structuring a complex idea or summarizing a long report. It keeps me focused, reduces mental clutter, and frees up space for the work that truly needs me. AI doesn’t replace thinking. I think it makes space for more thinking.

Haseeb Irfanullah

I don’t intentionally ‘use’ AI of any kind. I, however, can identify a couple of examples where I knowingly take AI’s help as it is a default part of a digital system: the ‘AI Overview’ section appearing after a Google search, and the manuscript processing platform of my journal recommending reviewers to me — that’s all!

This is not the first time I have avoided a new tool. I did the same when reference management software became commonplace two decades ago. After the initial refusal, I was less interested in learning about the tool and remained comfortable with manual referencing. Will the same thing happen to me with AI?

For AI, so far, my avoidance is a conscious personal choice. I enjoy communicating through my writing, slide decks, and spoken words. So, I put a lot of effort into preparing them (even if such elaborate effort is not really necessary). To me, it is like cooking: I know every ingredient, I know why it is there, what contribution it is making to the whole dish, as I savor it and enjoy others’ appreciation. My communication style could be another reason for my reluctance towards AI use. I often try to bring in ideas from apparently unrelated sectors and disciplines to create metaphors. I often try to see things through uncommon lenses. Maybe AI could help me to do these brilliantly, but I don’t want to know it! I adore my imperfections and mediocrity.

Also, as I get older, I want to keep my “little grey cells” active by doing things in a non-AI way, or nowadays we may call it doing things ‘manually’. I know that our mustaches are beyond compare, but I am sure my last reason would make mon ami Hercule Poirot happy!

Lisa Janicke Hinchliffe

For me, the right “hack” depends on the task at hand, and so I have a toolkit of approaches. Today, I will share three that support me in developing my thinking, creating instructional materials, and enhancing my writing. Before I go on, I also want to mention that I am using paid versions of the tools I mention here.

1. The Perspective Hack: My first hack is to use ChatGPT as an interlocutor. I input prompts such as “what questions should I be thinking about on XYZ topic?” or “what questions would ABC have about this topic?” This prompt helps me think around and through a topic from multiple perspectives and vantage points. ChatGPT is not giving the answers. Instead, generating questions that lead me to consider my own answers from multiple perspectives means ChatGPT is helping me think through my own ideas.
2. The Drafting Hack: I teach a lot of workshops. In any given week, I might be at the front of the room leading people through complex content 5-6 times. This is content that I know well, but like many people, I am often pressed for time to develop the training materials. I have found it helpful to bootstrap the development of session outlines and activities by providing ChatGPT with the topic, a description of attendees, and a list of learning outcomes, and then a prompt to “draft an outline for an n-minute workshop.” I make many edits to the output, but the efficiency gains are significant.
3. The Polishing Hack: I’m a bit late to this hack, but about a year ago, I started a subscription to Grammarly. I use the spellcheck and grammar suggestions, of course. But I have particularly appreciated the coaching on tone, sentiment, and conciseness. I also really enjoy the “Grammarly Insights” emails I receive that reveal patterns in my writing and areas for improvement. But, I have to say … telling me that I have “Unnecessary ellipsis — This mistake appeared in your writing at a higher rate compared to 51% of Grammarly users” … makes me laugh every time.

Stephanie Lovegrove Hansen

Marketing has consistently topped lists of the roles most likely to be affected by AI. From my vantage, that’s because marketing contains multitudes — when you do a little bit of everything, a little bit of each task can likely leverage AI. So though I have strong opinions about the portended automated future of Marketing teams, I have found myriad applications where AI has saved time, so I can focus on more high-value work. This includes: survey evaluation (surfacing key themes and areas of action), troubleshooting in web design (since we moonlight as pseudo developers), feedback on authored content (how can I make this better?), graphic design (removing backgrounds from images, extending graphic patterns for large-format designs), color alternatives for accessibility, and keyword extraction/assignment.

One of my favorite applications has been help in navigating the mind-numbingly unintuitive interface that is Google Analytics 4 (hot take: they did this on purpose to build a market for consultation services). I’ve also used it to extract email addresses from inconsistently formatted contact lists, to normalize the format and length of conference recaps from multiple contributors, and to do budget/location research for regional events.

Finally, monitoring brand cohesion for a growing organization is to play whack-a-mole with weird slide fonts, off-brand framing, and clearly AI-generated dross. I’ve been able to leverage our companywide Claude subscription and MCP capabilities to build voice/tone/style guidance directly into the tools so teams can filter content through these guidelines, improving consistency.

All these efforts free up time that I can then apply to more strategic, creative, and fulfilling efforts (and to writing strongly worded letters to Google).

Dianndra Roberts

AI, for me, has a place as an accessibility and accountability tool — Alexa, Gemini, and Co-Pilot have been useful when there isn’t necessarily a person available. Navigating work and life balance with Autism and ADHD can mean that I am often working with time blindness, now vs. not now thinking, and hyper-fixation. This means basic needs, such as food, hydration, and rest breaks, are ignored, for lack of a better word. It isn’t as simple as saying ‘stop’ when you’re hungry, or you’ll know when you need a comfort break, or just make a to-do list. There are many reasons our brains overlook these needs, and sometimes having a voice or a person in place can help to break the cycle.

Alexa, for example, can work alongside the Pomodoro method, create background noise/music, or make an announcement every so often to help me keep track. When body doubling might not be an option, these are alternatives that work for me, as well as reminders to check that I put the leftovers away from dinner, so I do not end up throwing out more spoiled food, or asking me what my water intake is for the day, because if I “forget to remember” then it will not happen, and ADHD tax will take its toll.

My phone and connected tools like Gemini set functional modes to help with potential distractions whilst focusing — locking down specific apps, timers for other apps, and using greyscale rather than colour screens during work hours and during nighttime for a better sleep routine. Sleep is important for us all, but there is an increased sleep disturbance for Autistics and other neurodivergent folks that ultimately impacts all aspects of our lives.

Co-Pilot can summarize long meeting threads, emails with extensive text, and hidden deadlines. Having a processing disorder, I might need more time to absorb large amounts of information, and sometimes my requests for bullet points, key dates in bold, or a TL;DR are missed. Using a summarizing tool helps me better understand and navigate what, if anything, is being asked of me, and to help collate the relevant information.

The growing use of AI does worry me, and it is becoming more ingrained in what we do. Of course, it has its benefits, and it would be amiss to say it hasn’t improved some work-life management for me. However, I am very aware of the environmental damage it causes and the use of AI to generate images, videos, repurposing memes, or as a search engine. Over the years, we have seen a rise in climate disasters. Far less often do Western societies feel the longer-term impact of the growing technological advancements, with other parts of the world left to pick up the pieces. So, if we are to use AI and it is here to stay, we should use it wisely.

Tim Vines

I get annoyed with AI assistants more often than I use them. For instance, Acrobat offering to summarize an invoice or a legal contract is just not helpful, as it’s the details that matter. Plus, the AI-based suggestions in my Apple Messages seem to have been hacked by my teenagers (“Of course I can pick you up from school at lunchtime!”). Google Docs is even jumping in and offering to “Help me write” as soon as I pause while writing this very post.

The AI-powered search in Notion and even in Google Drive can be shockingly bad, to the extent that I’m surprised it was released to the public. Perhaps I’ll take advantage of all these in the future, but right now I just want them all to go away. This sentiment applies doubly when the AI suggestion takes control of your cursor and you have to actively reject the suggestion (normally the opposite of what you want to say) before you can continue.

That said, the smarter autocomplete and next word suggestion in text messages is very welcome, as is AI-powered meeting transcription (although it’s annoying to have Google, Otter, and Zoom all offering their actionable summaries once the meeting ends). On documents that don’t need to be perfect, like applications for government funding or grant compliance reports. ChatGPT 5 does really well at taking in a range of relevant documents I’ve written and quickly making a draft of what I need that’s on topic and written in the same style. Since I’ve now got a paid subscription, it also has the great feature of remembering previous chats and being able to launch back into them when needed.

(NB: I also use DataSeer’s LLMs every day, but they’re deliberately set within strict infrastructure and focused on assessing data and data sharing in research articles rather than, e.g., compiling Christmas shopping lists, so I don’t use them to help with day-to-day office work.)