Diversity in the AI age – Panel discussion | Craft 2026

June 04, 2026

Diversity in the AI age – Panel discussion (Panel Outline)

A Craft 2026 panel discussion (moderated by Gabriella) on tech diversity — gender, age, LGBTQ and more — and the central question: is the rise of AI changing the picture, or are we just recreating the same patterns across the industry? Four panelists: Katalin (“Katia”) (software development manager at Tesco), Medea Vachkova (co-founder of the conference organizer team), Rosalia Miklos (“Rosie”) (head of the Budapest office at MSCI), and Jozsef Foldi (IBM sales, data & AI software; co-host of a Hungarian tech podcast). Outlined below as a topic-threaded conversation. Recurring thread: three of the four panelists (and the moderator) are parents of three, each with daughters, which colors the discussion of raising the next generation.

1. Opening & Framing

1.1 The premise

In technology, diversity ratios for women — and other dimensions (age, LGBTQ, etc.) — remain underrepresented, a long-term tendency visible even at the conference.
Core question: does AI change this, or do we recreate the same pattern?

2. Panelist Introductions & Why Diversity Matters to Them

2.1 Katalin (“Katia”) — Tesco

Software development manager at Tesco.
Does not believe in linear career progress — took multiple career breaks and returned at a lower level at different companies (e.g., from architect to a simpler developer).
Often the first part-time worker (or one of the first) at companies, specifically asking for it.
Mother of three — two girls and a boy — and a strong work-life-balance advocate.
On the moderator’s question (does having girls change your view?): “Mine doesn’t change at all.”

2.2 Medea Vachkova — conference organizer

Co-founder of the conference organizer team; joined the panel last-minute (decided “yesterday”) after realizing, as an expert listener, how much the company already does for diversity.
Concrete organizer practices from the very first event:
- Differential ticket pricing — if a boy and a girl come together, one gets a significantly lower price.
- Deliberate attention to speaker selection across “the whole spectrum,” not only women.
Mother of a seven-year-old son.

2.3 Rosalia Miklos (“Rosie”) — MSCI

Head of the Budapest office (~500 people) at MSCI, a New York-based global provider of investment-decision-support tools — a technology company in the financial domain, with a matrix structure.
Also holds a global role managing a large R&D team.
Long-standing personal cause: supporting women’s careers and women in STEM; MSCI has many employee resource groups (women in leadership, women in technology, all abilities, pride).
Firmly believes diversity is good for a company — different perspectives always serve it.
Mother of three — two boys and a girl (the girl the youngest, turning 3).
Reflection: she once thought you needed a daughter to influence future generations, but realized you have just as much influence raising boys — she and her husband model a family where both parents work and share logistics, which matters as much for boys to see as for a girl.

2.4 Jozsef Foldi — IBM + podcast

Two “hats”: (1) IBM Hungary sales, managing the data & AI software business for 10 years; (2) co-host of a Hungarian podcast (2.5 years, 40+ episodes).
Around episode 20 he realized all guests had been men — “I am now a minority representing the majority” — and decided to change it.
Had a good conversation with a contact from a technology university / HR leader organizing the “Top 50 Women in AI”, ran a campaign, and invited winners and finalists as guests.
Father of three daughters — the eldest just starting an IT university faculty.
Acts as an internal sponsor at IBM Hungary for more inclusion/diverse teams (the sales org had been all-male for five or six years).

3. Thread — Is AI Changing the Picture?

3.1 The paradox

Moderator: AI is democratizing knowledge (including coding), yet AI/AI-development meetups still show roughly 100 men to a handful of women — why?

3.2 Rosie — access was never the real problem; role models are

Logically AI democratizes access, but even before AI, gender-based lack of access to knowledge was already minimal.
The real problem is role models and mentors — someone telling you that you are empowered to do this.
Example: the STEM Sisters program pairs company mentors with high-school/university girls on tech projects.
A colleague’s mentee’s insight: working together and having help is nice, but the true value is seeing another woman already in a technological role at a company like MSCI, so she can walk in and feel 100% empowered — this is a human problem AI doesn’t solve.

4. Thread — It Starts Earlier: Education & Teachers

4.1 Katia — diversity is “late” by the workforce stage

Too few women in IT means we’re already late — it starts at university (few girls) and earlier, in education.

4.2 Katia’s personal experience

Her first year learning informatics in high school, the teacher said: ”Boys, come closer, you’re the ones I’ll teach — girls can’t learn this, so sit in the back row, I won’t explain it to you.”
When you’re young and a teacher says so, you believe it and act on it.
Better teachers later kept her in the field — the teacher has an enormous say.

4.3 It repeats 30 years later

Her daughter, in high school, had a similar experience: wanting to enter math/IT competitions, she wasn’t encouraged and had to convince teachers it was her place — it should be the other way around.

4.4 You can’t rely on luck — parent at home

Whether you get a good or bad teacher is luck; parents must encourage daughters and sons equally.
If you praise a son for something, praise a daughter for the same — and vice versa; this starts at home/parenting.

4.5 Moderator’s anecdote

His female math teacher asked his mother (a doctor) whether she’d done his homework — his mother said no (she understands her field, but not math).
Audience poll: hands up if you experienced comments that girls/women “don’t belong” in STEM — the moderator notes it seems widespread, and that early discouragement loses almost half the potential STEM workforce.

5. Thread — Concrete Practices That Work

5.1 Medea — female-speaker ratio as a must

For their conferences, a certain ratio of female presenters is a must.
Some male speakers refused to present unless there was at least 30% (ideally 50%) female speakers — they wouldn’t consider it a quality conference otherwise.

5.2 Medea — it’s hard to find female speakers

UX conferences never lack female speakers, but data and AI topics do — the team works “two or three times harder” to find and convince one.

5.3 Medea — accommodate mothers

Offer speakers/MCs the chance to bring spouse and kids.
Example: an MC (mother of a ~1.5–2-year-old boy) brought her son, who plays nearby; she engages with him during breaks then returns — the organizers provided circumstances to be both professional and mother.

5.4 Medea — “everything can be worked around”

Example (“Eva”): reduced hours after having a baby and kept them, staying both a professional and a mother.
Medea’s own routine with her seven-year-old: drives him to school, works from a nearby coffee place, picks him up first — no time lost in traffic, “the afternoon is ours.”
A workplace allowing this makes more women take similar jobs and aim higher, unafraid that a promotion costs family time — “it just takes a little flexibility.”

5.5 Katia — part-time by default at Tesco

Her boss advertised her position specifically as part-time because he wanted to hire a woman (Tesco Hungary had very few women).
She was one of the first part-time workers; it was such a success that every position is now part-time by default (candidates choose part- or full-time) — “a big achievement.”
But Central Europe is hard: ~10% women in IT here vs. ~30% in India — much more challenging in every sense.

5.6 Rosie — a family fun fact

Rosie returned to work after 6 months for most of her kids; her husband then stayed home and later job-hunted as a part-time employee with a small child, experiencing “pretty much everything mothers usually experience” (a company called Bay Kek Kah / BlackBird hired him — she gives them a shout-out).

5.7 Rosie — diverse interview panels

When hiring, “everyone wants to hire a mini-me” (“I see the young me in you”) — we engage with people similar to us.
In tech, an all-male, similar-background panel can pass on a candidate in the first round just because of this effect.
Diverse interview panels (different nationalities, genders, backgrounds) give a more diverse, reliable perspective for independently rating candidates — “worked for us very well.”

5.8 Jozsef — IBM sales vs. shared service center

IBM sales org in Hungary is ~90% male; the shared service center (~2,000 people) is almost the opposite.
HR’s explanation: sales doubles the barrier for women — not just hard technology, but an unfriendly work-life balance (client dinners, travel to conferences abroad/countryside), tough with kids.
The shared service center has a more usual schedule, hence the opposite ratio.
Hence his personal sponsorship to bring in more women — diverse sales teams perform better.

5.9 Moderator — the Women & AI meetup

The moderator started a meetup series, Women and AI.
A man asked why separate groups are needed / why women don’t just join the same groups.
Answer: it doesn’t happen without separate groups — women need a little extra help and a more comfortable environment where they aren’t immediately challenged by topic experts, learning step by step instead.

5.10 Katia — it’s about self-confidence

She’d rather attend a Women in AI meetup (and did, and recommends it).
The interview analogy: with 10 criteria, a man knowing 5 says “I’m the best, 9 out of 10” (overconfident); a woman who is genuinely 8 out of 10 downplays it (“not that bad… I know English, though only a C”).
Women prefer safer areas, then step slowly out of the comfort zone.

6. Thread — Parenting Both Directions

6.1 Katia — encourage daughters as you encourage sons

Praising a son for “challenging us all the time” as cool while telling daughters “girls don’t do that” plants a lack of self-confidence — encourage daughters to “be brave, do it.”
The reverse too: if a son wants to cook, don’t call it “a girly thing” — housework/chores are a good skill regardless of gender.

6.2 Rosie — it must work the other way around for boys

Tell boys they can cry, be kindergarten teachers, dance ballet — it’s not only about encouraging girls.
Her husband’s line: the only two things men can’t do are giving birth and breastfeeding; everything else they should be able to do.

6.3 Medea — Girls’ Day (Lányok Napja)

A Hungarian event twice a year: companies host in spring; a big autumn event gathers ~500 high-school girls interested in STEM.
The power is in creating a space only for them, where they feel allowed to do this.

6.4 Rosie — the positive-discrimination / impostor-syndrome dilemma

Counter-argument: as a US company under the Trump administration, there’s even a legal suit against Coca-Cola over a women-specific networking/development program — it’s not always clear you’re allowed to do that.
Risk: women-only quotas/events can be perceived as positive discrimination, planting the idea that women are only there because of it, not their capability — “a little dangerous.”
She personally has impostor syndrome: “am I where I am because of what I can do, or because the world now prefers a woman here?”
Important nuance: seeking women for roles/speaker slots is good, and there is no shortage of capable women — the real risk is planting the self-doubt, not filling seats with unqualified people.

7. Thread — Role Models in AI

7.1 Jozsef — the AI role models are all men

Naming the biggest AI role models — Sam Altman and others — they’re all men, usually CEOs (the top of the pyramid, with the resources and funds).
“That’s not good” — we must build up role models for women too.

7.2 Jozsef — limit positive discrimination to a threshold

Positive discrimination should be limited until reaching a certain percentage (unspecified), to encourage women out of the comfort zone to compete.
Panels/discussions like this matter for ordinary males who might not even see it as a problem.

8. Audience Q&A

8.1 Q1 — Does AI affect diversity in any way?

Jozsef: yes — bias. Training on an already-biased dataset that doesn’t represent the real world makes AI fail and give wrong decisions.
Hence IBM’s focus on AI governance and data governance; be careful how models are tuned; diverse datasets are crucial.

8.2 Q2 — Is female management different?

Medea: used to feel some women reach the top by ”acting like a man” (losing their woman attributes) — a pity; the deeper problem may be our belief that a leader must be very tough, far from how a woman would act. (Notes Rosie, next to her, is at the top and does not act that way.)
Rosie: told a male colleague that a “masculine aggressive management style” is annoying; he pushed back — maybe 1% of men and 0.5% of women act like that, yet we perceive “men are like this.” There’s no inherent male/female style; personalities differ. She’s often the only woman on a management team and finds she brings different topics/perspectives that get positive reactions (“good to have you in the room”) — but we over-generalize by gender.
Katia: far less difference between us than we think; good leaders are leaders, not bosses, with a human perspective. Reaching a leadership position tends to correlate with being either a narcissist or someone who puts people in focus — independent of gender; she’s seen very negative female leaders and very human male leaders.
Jozsef: never had a female manager until recently — a new regional manager from Croatia (a senior executive he’s known 10+ years), “a lady kind of manager” bringing different ideas (e.g., different sales-cadence calls) — he already sees the change.

9. Closing

9.1 Moderator’s summing-up

More similarities than differences, but women often bear responsibility for family and their wider environment, so they naturally bring a perspective of cooperation, not just competitive edge.

9.2 The pledge

Audience asked to promise to do something in the next week: spread the word that there are great female leaders in IT and AI, and help them flourish.

People & References Cited

Gabriella — panel moderator (also referenced as connecting to Jozsef’s podcast).
Katalin (“Katia”) — software development manager, Tesco; work-life-balance / part-time advocate.
Medea Vachkova — co-founder of the conference organizer team.
Rosalia Miklos (“Rosie”) — head of MSCI’s Budapest office; global R&D leader.
Jozsef Foldi — IBM Hungary (data & AI software sales); co-host of a Hungarian tech podcast.
“Eva” — cited as an example of reducing hours after a baby while remaining a professional.
Sam Altman — cited as a (male) example AI role model / CEO.
Companies & orgs: Tesco (part-time-by-default), MSCI (investment-decision tools; employee resource groups), IBM (Hungary sales; ~2,000-person shared service center; AI/data governance), Bay Kek Kah/BlackBird (hired Rosie’s husband as a part-time parent), Coca-Cola (US legal suit over a women-specific program).
Programs/initiatives: STEM Sisters (company mentors + girls), “Top 50 Women in AI” campaign, Women and AI meetup, Girls’ Day (Lányok Napja).
Concepts: role models over access; bias starting in education/parenting; teacher influence; part-time-by-default; diverse interview panels (“mini-me” hiring bias); women-only spaces and self-confidence; the interview-criteria confidence gap; positive discrimination vs. impostor syndrome; AI training-data / dataset bias and governance; whether a distinct “female management style” exists; leaders vs. bosses; regional ratios (~10% Central Europe vs. ~30% India).

Video: https://www.youtube.com/watch?v=uiExKhnhSt8 — Transcript via yt-transcript.sh; outline generated from the transcript.