BozarthZone

Ebola! An ethical dilemma -- and extending the reach of L&D

2026-08-31T10:15:09.636-04:00

About ways to extend the reach of L&D, a true-life ethical dilemma, and when AI can be a help -- or maybe not.

Spoiler: We brought them home and all of them survived.

[PS: I meant to say "NOT a black & white issue".]

hashtagLandD hashtagID hashtagReflectivePractice hashtagSYW hashtagShowYourWork hashtagWOL

Video: Woman wearing blue floral shirt, who has reddish hair and glasses, seated in office.

https://youtu.be/yr5cm_UkeYQ

The Green Milk Incident: When AI Lacks Context

2026-08-27T09:37:01.665-04:00

Subject: Food-tech firm NotCo and its AI system "Giuseppe", working on development of plant-based milk.

"The milk came out green. Giuseppe had detected properties in dill that appeared useful in the data, but it missed something obvious to anyone standing in NotCo’s kitchen: whatever else milk may be, people do not expect it to be green. The failure made it obvious to everyone involved that—despite having access to large amounts of data—Giuseppe had a serious lack of context.

The green milk vividly illustrates the different nature of machine knowledge. Giuseppe excels at detecting potentially useful relationships among ingredients and molecular compounds. But it cannot taste the product, smell it, or understand everything people expect from a glass of milk. In contrast, chefs can recognize if a prototype is too grassy, too thin, too sweet, or simply wrong. They can appreciate color, aroma, mouthfeel, aftertaste, stability, and performance under heat. Much of this knowledge has been acquired through experience and is difficult to reduce to rules."

The Dark Side of Community

2026-08-18T10:30:00.000-04:00

We talk a lot about the benefits of community at work.

Communities of practice. Networks. Peer learning. Knowledge sharing.
Many of those conversations start with a pretty optimistic assumption:
People want to connect, participate, share, and help each other, and
usually they do.

But here's the thing I've learned from decades spent looking closely at communities:

Community isn't automatically a good thing.

https://youtu.be/BQMyHBQZhdI

Stay Balanced!

2026-08-14T09:37:09.886-04:00

Egad - another friend joined the "Women With Broken Wrists" club yesterday. Since there was interest in my earlier post about the balance class I took I thought I'd offer a few takeaways.

-Change up while walking: Do some figure 8s and zigzags, practice hard sudden stops, walk backward a few feet.
-Practice tandem standing (feet heel to toe) and standing on one leg. Look in other directions, close your eyes, etc. The biggest help I found with this was practicing on a thick (2") stability pad like they use in yoga and PT.
-The Tai Chi session recommended beginner videos by Dr. Lam (below).
-The instructor who led the practice in falling and getting up session (getting up is not easy if you've a broken wrist...) recommended the "Bob and Brad" videos below.

Dr Lam Tai Chi for Beginners Bob & Brad: How to Fall Bob & Brad: Getting Up Off the Floor

https://unsplash.com/photos/photo-of-silhouette-photo-of-man-standing-on-rock-U3C79SeHa7k?utm_source=unsplash&utm_medium=referral&utm_content=creditShareLink

Do you even know what help you need?

2026-08-10T09:36:00.000-04:00

Something different: I have been in conversation lately with several friends who are struggling with living their own lives and doing their own jobs while trying to manage medical situations for their aging parents. Nearly everything takes a lot of time, from sitting in waiting rooms to being on hold with insurance companies.

I had a plate full of this some years back when my husband had very complex surgery and needed a lot of aftercare while I was still working full-time. Honestly, I felt like I was drowning. IYKYK.

Did you know you can hire a part-time helper with the job title “Medical Manager” (edit: aka "patient advocate"). Changed our lives. She excelled at all the things I found infuriating and exhausting: wrangling with insurance companies, schedulers at doctors’ offices, pharmacies, etc., and even sometimes taking him to appointments or being the required driver available to take him home after procedures. Charged by the hour and she wasn’t even very expensive. Worth every penny. There might even be some insurance coverage for that.

Hope that helps.

Women Talking About Informal Learning

2026-08-04T15:08:52.622-04:00

Fun with Nahdia Khan on the latest "Women Talking About Learning" podcast.

"Discover the transformative power of informal learning in the workplace! Join Jane Bozarth and Nahdia Khan as they chat about how informal learning can shape professional development. Don't miss this insightful episode! Discover the transformative power of informal learning in the workplace! Join Jane Bozarth and Nahdia Khan as they chat about how informal learning can shape professional development. Don't miss this insightful episode!"

"The Informal Learning One"

About electricity-- I mean, AI

2026-07-23T08:46:05.703-04:00

Using Your Business Levers (ATD Blog Post)

2026-07-07T10:25:40.798-04:00

So much of the talk around L&D focuses on things that the an indivdual practitioner just doesn't control: C-Suite strategy, change initiatives, budget, staffing, etc. But we do have areas where we can exert a lot of influence. Use your levers!

https://www.td.org/content/atd-blog/show-your-value-by-working-your-business-levers

Mapping Discussion Flow Tool

2026-07-06T10:11:36.705-04:00

Explainer: "Mapping Discussion Flow".

Great feedback for new (and experienced) presenters, trainers, and facilitators.

This provides an interesting snapshot of a segment of a learning experience and opens the door for conversation about developing facilitation skills. Participants in my train-the-trainer classes loved this; it's easy to learn to do and they often volunteered to map each other in the leadup to their final teachbacks.

Also: A fabulous artifact for showing & sharing your work. #syw

Captioned video: https://youtu.be/t98BMrWCOLA

Written directions available at https://lnkd.in/eVw5mVwg

What does your organization notice?

2026-07-02T11:48:55.139-04:00

I was in a conversation yesterday about how individual practitioners can show expertise.

What do people pay you for that never appears in your job description? (And what did I miss?)

Which of these does your organization notice?

Relationships
Network awareness
Political navigation
Reading the room
Coaching
Diagnosing root causes
Building trust
Connecting people
Facilitating difficult conversations
Recognizing patterns
Knowing when not to create training

LearningNow TV features "Adaptive Enterprise Framework"

2026-06-27T12:31:00.000-04:00

Check Conversation about the Adaptive Enterprise Framework on LearningNow TV!

Fun conversation with Kate Graham; Episode June 25, 2028. PDF available for download at the same link.

https://learningnow.tv/watch/programme-145---25-june-2026

Image: Jane Bozarth speaking.

Fun New "Show Your Work: Podcast Series with Andrew Jacobs

2026-06-11T12:36:00.000-04:00

"Women Talking About Learning" podcaster Andrew Jacobs and I are currently recording a fun 6-episode series called "Show Your Work", taking on live unplanned conversations about informal learning, social learning, problems with evaluation, working effectively with others who disregard evidence, and Like That.

Schedule TBA, but sometime soon.

Image: banner ad, "Bozarth & Jacobs: Show Your Work".

Notes on the Canadian eLearning Conference 2026

2026-06-03T12:49:00.000-04:00

Notes on the fun and inspiring Canadian eLearning Conference, with shoutouts to some great participants.

About Vanity Metrics...

2026-05-14T12:46:00.000-04:00

Recent news about Amazon shutting down a token AI Leaderboard ("don't use AI just to use AI") reminds me of this old chart from Jennifer Laycock, formerly of Search Engine Guide.

Vanity metrics don't necessarily = desirable outcomes.

2026-04-21T12:10:00.000-04:00

𝐂𝐨𝐥𝐥𝐢𝐬𝐢𝐨𝐧 𝐒𝐩𝐚𝐜𝐞𝐬: 𝐑𝐞𝐝𝐮𝐜𝐢𝐧𝐠 𝐒𝐨𝐜𝐢𝐚𝐥 𝐅𝐫𝐢𝐜𝐭𝐢𝐨𝐧 𝐢𝐧 𝐃𝐢𝐬𝐭𝐫𝐢𝐛𝐮𝐭𝐞𝐝 𝐖𝐨𝐫𝐤

A casualty of the shift to remote and hybrid work was the removal of many natural collision spaces. It's less common now to see places workers used to just run into each other, as with the literal water cooler, break room, or even stairwell. These moments helped people see what was happening elsewhere, meet new colleagues, discover unexpected expertise, and build a sense of “we’re in this together.” Such "collisions" can go a long way toward providing a psychologically safe environment, let people connect as peers, and help support a healthy culture.

While onsite workers may have now have access to high-traffic spaces intentionally designed to break down silos and encourage casual interaction -- like coffee bars or lounge areas with whiteboards -- distributed workers, don't encounter those collisions, and most digital environments don’t replace them.

𝐒𝐮𝐠𝐠𝐞𝐬𝐭𝐢𝐨𝐧𝐬 𝐟𝐨𝐫 𝐜𝐫𝐞𝐚𝐭𝐢𝐧𝐠 𝐚𝐧𝐝 𝐬𝐮𝐩𝐩𝐨𝐫𝐭𝐢𝐧𝐠 𝐦𝐨𝐫𝐞 𝐝𝐢𝐠𝐢𝐭𝐚𝐥 𝐜𝐨𝐥𝐥𝐢𝐬𝐢𝐨𝐧 𝐬𝐩𝐚𝐜𝐞𝐬:
Think about virtual spaces that mimic the casual interaction of the office,such as Slack channels for non-work topics (e.g., pets, hobbies), virtual coffee chats, or digital, interest-based groups.

Create or enhance these opportunities to support performance and the work culture and to alleviate the feelings of disconnection experienced by many remote workers.

𝐀𝐧𝐝 𝐠𝐞𝐭 𝐀𝐈 𝐭𝐨 𝐡𝐞𝐥𝐩 𝐲𝐨𝐮:
AI can help compensate for the absence of natural collision spaces. It can:
--suggest relevant conversations or communities based on work and interests
--surface shared challenges, complementary skills, or overlapping goals
--summarize meetings and threads so people can step into ongoing conversations
--create continuity across asynchronous exchanges
--create dynamic "office maps" where presence signals invite drop-ins
--house persistent digital whiteboards that connect ideas across sessions and teams
--summarize asynchronous conversations so people can more comfortably jump in mid-conversation

It's important to understand that this is not some sort of scheduled collaboration but, rather, a matter of restoring the conditions for more serendipitous, spontaneous connection, as with that water cooler. Giving space and places for safe opt-in, asynchronous (across time zones) silo-spanning connection that requires little activation/energy can help workers feel supported as well as provide catalysts for ideas and innovation, increase tacit knowledge sharing, and provide a peripheral awareness of sharing a physical space within the organization.

Extra credit: If you're interested in a more academic view, read up on "relational density" and its relationship to trust.

What ideas do you have for creating or supporting collision spaces for your remote and hybrid workers? What are you seeing work to create these kinds of collisions in distributed teams?

2026-04-07T16:07:00.004-04:00

What do humans bring? (6)

Social infrastructure to continuous learning to human meaning-making: AI is an AMPLIFIER, not a driver, of these things. It can find patterns, uncover redundancies, and speed up access, output, and experimentation.

But in many organizations it can't do that without the humans.

https://youtu.be/YtUw6lAZ9Xg

2026-04-06T08:06:00.002-04:00

𝐖𝐡𝐚𝐭 𝐝𝐨 𝐡𝐮𝐦𝐚𝐧𝐬 𝐛𝐫𝐢𝐧𝐠?(5) 𝐌𝐨𝐫𝐞 𝐨𝐧 𝐡𝐮𝐦𝐚𝐧 𝐦𝐞𝐚𝐧𝐢𝐧𝐠-𝐦𝐚𝐤𝐢𝐧𝐠

I posted about this on Friday. Social infrastructure and continuous learning are brought to bear in human meaning-making: How people interpret, question, and apply what they know. This demands contextual, rather than procedural, decision-making, involves weighing competing "goods" (like efficiency vs. safety) rather than a clear right or wrong, and often requires social awareness. And: it's frequently invisible work, where decisions are hard to see but impact can be huge.

Consider: If these decisions are left to AI, who is accountable if something goes wrong? AI will not be fined, fired, go to jail, feel guilt, or stay up at night worrying.

Here are some examples of human meaning-making. What others can you think of?

𝐒𝐚𝐟𝐞𝐭𝐲
When to shut down operations due to risk vs. continue under pressure to meet deadlines
Interpreting near-misses: anomaly or warning sign?
Adapting rules to real-world conditions (weather, fatigue, equipment variability)
Reporting safety issues that could delay projects or cost jobs
Enforcing rules consistently vs. making exceptions for experienced workers
Balancing productivity metrics with human well-being
Recognizing when a “compliant” situation is still unsafe

𝐂𝐨𝐧𝐬𝐭𝐫𝐮𝐜𝐭𝐢𝐨𝐧
Adjusting plans when site conditions don’t match designs
Deciding whether to proceed with imperfect materials or wait (cost vs. quality)
Pressure to cut corners vs. long-term structural integrity
Navigating “this is how we’ve always done it” vs. safer/better methods
Making tradeoffs between craftsmanship and speed

𝐇𝐞𝐚𝐥𝐭𝐡𝐜𝐚𝐫𝐞
Choosing between treatment options with different risks
Prioritizing patients when resources are limited
Using AI recommendations vs. overriding them
Managing uncertainty—when evidence is inconclusive
Balancing protocol adherence with individualized care
Communicating difficult news with empathy

𝐋&𝐃
Choosing when to push standardized training vs. allowing informal/social learning to emerge
Interpreting incomplete data from learning analytics (for example: low completion ≠ low capability)
Using AI to personalize learning vs. protecting employee privacy
Deciding if a “good enough” solution is acceptable under time pressure

2026-04-03T08:46:00.001-04:00

What Do Humans Bring? (4): Human Meaning-Making

I submit that the next layer of "what do humans bring to organizations?" is meaning-making. What do they do with what happens in the social infrastructure and continuous learning layers?

This includes things like:
How people interpret, question, and decide
How they determine what matters, what to trust, and what to act on
Getting beyond information and "content" to judgment, context, and ethics shaping outcomes

(Note that the word "layer" is intentional. I am not thinking of a hierarchy, steps, or stack, but a dynamic, interdependent, reinforcing system. I am trying, in part, to get beyond the ages-old problem of "learning" being viewed as an event.)

2026-04-01T11:32:00.003-04:00

What Do Humans Bring (3)? Continuous Learning

Connie Malamed told me to keep going...so I am. So far this week, in pondering what humans bring to the workplace, I've talked about my ideas around social infrastructure (made up of networks, Communities of Practice, working out loud/showing your work, and culture), and the ways L&D can support the humans in those spaces.

As I see it, the larger picture of organizational learning can be fleshed out into what I think of as the "Continuous Learning" layer, where the activities and learning in the social infrastructure are supported by formal, more structured things: LMSs, LRSs, traditional formal instruction, and the like. (I can't decide where to put performance management systems, but I'm thinking here, maybe?). It's not another tier --this isn't a hierarchy-- but another piece that completes the idea of "continuous learning" that happens in the organization.

And social learning happens here as well, as people -- we hope -- engage in robust, 𝘴𝘬𝘪𝘭𝘭𝘧𝘶𝘭𝘭𝘺 𝘧𝘢𝘤𝘪𝘭𝘪𝘵𝘢𝘵𝘦𝘥 (I repeat: skillfully facilitated) learning experiences built on realistic scenarios, real-world practice, and applicable takeaways, in settings that allow them to make meaningful connections.

In short:
Continuous learning (formal layer) = designed, structured, intentional
Social infrastructure = emergent, relational, adaptive

2026-03-31T11:26:00.010-04:00

What Do Humans Bring? (2) More on Social Infrastructure

More: The pillars of Social Infrastructure are networks, communities of practice, working out loud/showing your work, and organizational culture & its micro-cultures.

Where does L&D fit in here?

L&D’s role here isn’t to push out more content, but to shape how work, learning, and decision-making actually happen across the organization. That means
-helping communities of practice (CoPs) stay active, in a space where people working to improve their practice can share real work, compare approaches, and learn from each other in context.
-making it easier to surface what people are trying, what’s working (and what isn’t) so knowledge grows instead of getting lost or duplicated.
-introducing simple habits like debriefs so groups pause, reflect, and improve over time.
-helping managers create space for transparency, honest conversation, and experimentation.
-helping connect people across teams and networks so ideas don’t stay siloed.

In short, L&D supports the social infrastructure by strengthening the everyday patterns through which people learn, share, and get better at their work together.

AI: Who Checks the Checkers?

2026-03-31T10:59:00.000-04:00

Are those "tells" as telling as we think?

I do use an agent for editing and tightening up my writing. Then found out AI checkers are flagging content that is...(checks notes).... "clear and well-structured". Ok then.

Jane: "AI checkers say this is mostly AI-written, even though I wrote it; it uses my style and voice and even my preferred words like 'surface' and 'siloed'. Why does it think I didn't write it?"

Claude: "Short answer: those detectors are guessing based on patterns, not actually knowing who wrote it—and your writing happens to share some of those patterns.
A few things are working against you here:
Your writing is clear and well-structured
AI detectors tend to flag text that is:
logically organized
evenly paced
grammatically clean
low in typos or false starts
That’s…also what experienced professionals write like. Especially someone used to publishing."

I am reminded of the day I was accidentally copied in on an email in which work submitted by one of the most straight-arrow people in our business was accused of submitting "AI generated" content for publication. It was material adapted from a course they'd been teaching for years.

2026-03-30T11:21:00.000-04:00

What Do Humans Bring (1)? Social Infrastructure

Cogitating, ruminating, showing my work. What are your thoughts about "Social Infrastructure" and what humans bring to the workplace?

https://youtu.be/g--T7k1GT74

2026-03-26T10:08:00.004-04:00

𝐘𝐨𝐮 𝐂𝐚𝐧’𝐭 𝐀𝐮𝐭𝐨𝐦𝐚𝐭𝐞 𝐓𝐫𝐮𝐬𝐭

I’ve written a lot over the past few weeks about ways AI can support Networks, CoPs, and Showing Your Work/Working Out Loud. It can do a lot of things to amplify our efforts, corral our messy work habits, and generate rivers of content.

But AI can’t:

♦create trust
♦generate reciprocity
♦produce psychological safety
♦substitute for shared experience

𝘊𝘶𝘭𝘵𝘶𝘳𝘦 is where those things happen, in organizations where people are allowed to talk with each other, given opportunities to get together, enabled to voice concerns and objections, and can admit problems or mistakes without fear of the sky falling. That's a culture that fosters trust, reciprocity, inclusion, and psychological safety, values shared experience, and provides “collision spaces” like break rooms and water-cooler corners, both literal and virtual. Culture is about transparency. Culture permits the existence and effective use of the things I’ve been talking about for the past 20 years: strong networks, communities, CoPs, and showing your work/working out loud.

One—ok, two-- of my lasting concerns with conversations regarding “culture” is that the word itself is devilishly hard to define, and often it feels like the onus for creating an effective/productive/supportive culture is on management alone. Workers may not create enterprise-wide culture, but we absolutely help shape local culture. I mean, every team and community and department has its own culture, right? Its own norms and ways of working? (They’re called micro-cultures, btw.)

What culture(s) are you a part of? What can you, as an individual, do to help shape it? As an L&D practitioner, what influence might you have over it?

2026-03-24T09:10:00.003-04:00

Assessing the Value of Online Interactions

Note: This originally appeared as a "Nuts and Bolts" column in Learning Solutions Magazine, October 2012.

A good deal of my time is spent providing workshops and conference presentations on social learning and the use of social media to support and extend social learning in the workplace. In every session, it seems, someone comes just to challenge me to “prove” that all this isn’t a waste of time, that there is performance-enhancing value in social connections and interactions, particularly of the online variety.

They usually want some magic metric, some formula like, “two hours on LinkedIn + four comments in groups = tangible outcomes for the organization.” It doesn’t work that way. A great deal depends on how the worker chooses to spend that time in social channels, how well he filters and curates information, how she chooses the people with whom she’s interacting. The quality of those interactions depends in turn on many other issues, including trust, a willingness to ask for and offer help, and time invested in developing ties deeper than those purely at the surface. Likewise, a worker expected to improve performance and support organizational goals must know what the expectations are around that.

Value creation

Etienne Wenger (of CultivatingCommunities of Practice fame), Beverly Traynor, and Maarten De Laat have recently published a new conceptual framework for understanding and assessing value in such interactions. It includes a nice overview chart (figure 1) that I’ve found helpful in addressing concerns of my audience members.

Figure 1: Wenger, E., B. Traynor, and M. De Laat. Chart from Assessing Value Creation for Communities of Practice and Networks: A Conceptual Framework. Used with permission.

Immediate value

I’ll use myself as an example of how the chart helps shine light on real activity and outcomes. I spend a lot of time on Twitter because there are so very many smart people there, who at any hour of the day or night are talking about something I often didn’t even know I wanted to talk about.

I mostly follow learning, training, and eLearning people, but I also like some fiction authors and a few experts in other fields. Those people who only talk about what their cats had for breakfast? I don’t follow them. But it’s important to note: I am very active on Twitter. I engage, and talk, and interact with people. I drop in on several live Twitter chats a month. I try to contribute as much as I take. I like to think I help. So in looking at Wenger et al’s first column: I feel I get immediate value from the quality of interaction and reciprocity, I am given food for thought that I do reflect on, and I make it no secret that I am having fun.

Potential value

Moving across the chart to the second column: From my participation, what is the potential value? I’ve certainly developed a lot of connections, many in other parts of the world who offer very diverse viewpoints. I find I’m often inspired to read up on a new area or check out a new app or other tool.My views on learning have shifted considerably over the past five years as I’ve recognized firsthand the power and potential of increased support for social learning in the workplace.

Applied value

Now, moving to the third column, we look to see whether dots are connecting. I spend a lot of time on Twitter, I make a lot of connections,I read about things that interest me. But am I getting applied value? Do I leverage those connections? Have I engaged enough with my personal learning network so that, if I ask for help, some people might respond?

Let’s revisit an example I used in a previous column, one spurred by a phone call from one of our agencies.

I tweeted this (Figure 2):

Figure 2: Leveraging connections on Twitter: original tweet asking for help

In two minutes’ time I had several responses, including this one (Figure 3):

Figure 3: One of the many immediate responses

I found the document, scanned it to see if it seemed okay, and sent it on to the agency. They said it was just what they needed. This amounted to a four-minute interruption in my day.

So you tell me: Is there applied value? Am I using my connections and implementing advice?

Realized value

Moving to the next column on the chart from Wenger et al, “Realized value.” I gave the customer a good response in four minutes. Is that a reflection on my personal performance? How about my organization’s reputation? Let me ask it another way: when’s the last time you called a government agency and got a good answer in four minutes?

Reframing value

In terms of the last column of Figure 1, “Reframing value”: I don’t know that I’ve changed my institution (yet), but I’ve influenced ideas around new ways of working. And while I’m not asked for evidence that I am effective, whenever I get a solution or innovative idea via one of my social channels, I take a screenshot or write a quick note and send it on to management anyway.

So, in looking for value in online interactions, try to get past the idea of a magic metric. I can’t tell you that my spending x hours on LinkedIn and tweeting y times per day will get you the result I got in the example above. I can tell you that my choice of when, with whom, and how to engage is what helped drive that result.

What can we do?

So what can we do? Help workers begin to articulate the ways in which interactions have solved a problem, reflected on their personal performance, or reflected on the organization’s reputation or performance. Start asking, “What did you learn today/this week? How has that affected your performance? How does it help the organization?” Help connect dots between social interaction and access to expertise, and between those connections and new tools and reframing ways of working. And please do review the full text of the piece by Wenger, Traynor, and De Laat, available at https://www.betterevaluation.org/sites/default/files/Wenger_Trayner_DeLaat_Value_creation.pdf .

2026-03-23T10:25:00.001-04:00

"𝐖𝐞 𝐰𝐞𝐫𝐞 𝐠𝐨𝐨𝐝 𝐚𝐭 𝐭𝐞𝐚𝐜𝐡𝐢𝐧𝐠 𝐩𝐞𝐨𝐩𝐥𝐞 𝐭𝐨 𝐩𝐫𝐞𝐭𝐞𝐧𝐝 𝐭𝐨 𝐥𝐢𝐬𝐭𝐞𝐧."

Years ago I was a facilitator for a highly structured commercial management training program, essentially a discussion planning system. The first "Key Principle" was "listen and respond with empathy." Participants scripted and rehearsed and delivered on the principle, usually pretty convincingly, before moving on to telling the practice employee they were in trouble, or being laid off, or would have to start working Saturdays...

I remember back then a certain L&D author writing, "We are very good at teaching people to pretend to listen," and thought at the time, "We are good at teaching people to pretend to be empathetic."

Can empathy be taught? To AI? There are lots of research reports looking at what aspects of human behavior/communication AI can successfully be trained on. Here are 2, one from Ziao et al. and the other from Kumar et al., that crossed my path this week. Ziao et al. explicitly discuss using LLMs to teach "empathy"; Xiao et al., which I see as related, focus on whether AI can move beyond tasks to deeper aspects of teamwork like collaboration and "fragile communication".

My interest is in how empathy is defined. Xiao et al.’s study participants view empathy as an inner social state that is difficult for machines to replicate. Kumar et al. define empathy for their study as a set of linguistic strategies (prescriptive behaviors like validating emotions).

What do you think?

Kumar, A., Poungpeth, N., Yang, D., Lambert, B., & Groh, M. (2026). Practicing with Language Models Cultivates Human Empathic Communication. arXiv preprint https://arxiv.org/pdf/2603.15245

Xiao, Q., Hu, X. E., Whiting, M. E., Karunakaran, A., Shen, H., & Cao, H. (2025). AI Hasn't Fixed Teamwork, But It Shifted Collaborative Culture: A Longitudinal Study in a Project-Based Software Development Organization (2023-2025). https://arxiv.org/pdf/2509.10956