Exhibit B - My Departure: Dispatches from Inside DSIT's Central AI Risk (Dys)Function: How Hostile Leadership Broke Its Own Mission
They awarded me a bonus for my "brilliant work ethic". After I raised legitimate concerns upon my resignation, they erased me like I never existed.
**For the avoidance of doubt: I am not suicidal. I am young, in excellent physical and mental health, in full possession of my mental faculties and am looking forward to building a free, full life. Any sudden deterioration, unexplained harm to my person, my wellbeing or any unexplained interference with my communication must be treated with the highest suspicion.**
Hello. Welcome to Day 4.
I know that yesterday I said that I would be sharing further corroboration of the hostile environment in CAIRF from a second colleague, but I realised it would make more sense to share these pieces of evidence first.
As I said before, this is my archive. I’m building it however I see fit. It will include evidence of institutional abuse, betrayal and avoidance from all parties named in my initial piece. I will continue to publish daily (that is - until the survival instincts of those responsible finally kicks in and they decide to resolve this).
But, since they haven’t shown any signs whatsoever of wanting this to end, I’ll continue to publish.
At this stage I should also make one thing very clear: I am not out here leaking state secrets. I am not violating the Official Secrets Act. This isn’t WikiLeaks. This isn’t whistleblowing á la Edward Snowden’s NSA files.
I am bearing witness to my lived experience of discrimination, institutional abuse, ethical decay and what all of this means for our shared future. I’m articulating the hypocrisy. I’m also archiving the professional and moral failure of supposed “advocacy” organisations whose raison d'etre is to intervene to protect people like me.
Today’s piece will detail what happened around my departure from CAIRF.
I. The Bonus
On 8th November 2024, I was nominated by Dean Whitehouse, CAIRF’s Head of AI Risk Assessment, for an In-Year Award for the value of £150. In his nomination, Dean wrote:
“… joined the team recently and was immediately seconded from engagement to assessment. She threw herself into the challenge, delivering a number of excellent mapping outputs for the risk register which gathered compliments from the wider team of the usefulness and quality of these. She has demonstrated a brilliant work ethic, pushing herself outside of her comfort zone and delivering high quality and comprehensive results. She has further been working on delivering causal mapping, being pro-active in setting out a clear approach, working across teams and learning non-stop from her peers.”
Hold Dean’s words about me in your mind as you read what happened next.
II. My Resignation
My formal notice of resignation was sent on 9th January 2025. It was addressed to my line manager, CAIRF’s Head of Engagement and Preparedness, and Dean Whitehouse, CAIRF’s Head of Risk Assessment. I have reproduced my resignation letter below:
I am writing to formally resign from my role as an AI Risk Advisor with the Department for Science,Innovation and Technology, effective immediately. This decision was not made lightly, and I deeply appreciate the opportunities and support I have received during my time in this role.
I understand my contract requires a notice period, but due to personal circumstances, I am unable to serve it and must resign with immediate effect. I hope this can be understood, and I am happy to assist in handing over any responsibilities remotely.
Working within CAIRF has been an immense privilege, and I have learnt a great deal from the important work we do in safeguarding against the risks associated with AI. I am truly grateful for the professional growth I have experienced and the chance to contribute to this critical mission.
However, as I depart, I feel it is important to share some reflections. Throughout my tenure, I have observed and experienced behaviours and attitudes within the team that I believe warrant attention. While one of our core focuses is addressing AI biases, I feel it is equally important for us, as a team, to reflect on and address any biases within our own culture.
At times, I have felt out of place within what has often felt like a male-centric environment. Whilst I have the utmost respect for the expertise and competence of my colleagues, the normalisation of this dynamic can create barriers for those who do not fit the dominant culture.
This dynamic is not only personally disheartening but also counterproductive to CAIRF’s mission. A team that actively reflects on and works to address its own biases will not only foster inclusivity but also strengthen its credibility and effectiveness.
I share this feedback with respect and hope that it can serve as a constructive prompt for introspection and positive change. I firmly believe in CAIRF’s potential to lead by example, and I hope that future efforts will include a closer examination of internal dynamics
….
Thank you again for the opportunity to contribute to this team’s vital work. I wish you all continued success in safeguarding against AI risks.
Despite everything they put me through, my resignation letter was measured and respectful. I still believed, as I always have, that professionalism matters.
I did not resign lightly. It was absolutely heartbreaking for me to have to walk away from my dream role. In my letter, I cited ‘personal circumstances’ as the reason for my immediate departure. Let me be clear: those personal circumstances were the cumulative harm caused by the constant undermining and sabotage, the public belittlement after returning from sick leave for depression, and the lack of a safe working environment.
I resigned as an act of self-preservation. Staying would have meant sacrificing my health, my dignity, and my humanity for a team that had already decided I was “useless” and disposable.
III. My Erasure - The First Instance of Retaliation
Within days of my resignation, Dean erased me from all Microsoft Teams groups - without ever acknowledging my departure.
He remembered to erase me from the digital spaces I was a part of. He did not remember to acknowledge my humanity. My existence.
This was the first instance of retaliation I endured. This wasn't just procedural - it was deliberate. Personal. A way to dismiss me, shrink me, and signal that my presence never even mattered in the first place.
I was polite to Dean. I went out of my way to congratulate him when he got engaged. But he mistook my kindness for weakness. Fatal mistake.
Now, couple this with Dean’s words about me and my work from my bonus nomination - “brilliant work ethic”, “excellent output”, “pro-active”, “high-quality and comprehensive results”…
Judge for yourselves.
IV. Reflections
CAIRF’s Head of Engagement and Preparedness had acknowledged the sick note from my doctor stating my diagnosis of depression back in November 2024, writing this to me:
I’m sorry to hear you’re not feeling yourself, let me know if me or the support networks in DSIT can be of help in any way.
I appreciate you will not be signed in to your work IT but when you get a chance, please let me know how you would like to keep in touch so we know how we can continue to support you.
My work mobile number is in this signature so please drop me a message if you want to discuss anything.
He was aware of my diagnosis. He offered “support” when I was diagnosed. When I resigned, he gave me silence. This was deeply dehumanising.
This is what “AI ethics leadership” looks like at DSIT. These are the same people who are responsible for leading the government’s efforts to mitigate AI harms.
These are the people who DSIT senior leadership chose to protect. Not just protect, but elevate. Promote as the department’s face, voice and authority on AI safety and ethics.
In March 2025, Dean Whitehouse was put on a literal stage to represent DSIT, lecturing about the factors that contribute to a lack of trust, including concerns about safety, bias. In October 2025, he will sit as a judge at the 2025 Robotics & Automation Awards. According to event’s official website, Dean’s biography describes him as having “a proven track record of building and leading high-performing teams”.
Read that again.
Now re-read what it was like to work under his leadership - the hostile environment, the pettiness, the public humiliation, the gendered exclusion, the retaliation when I spoke up. Read the corroboration of the hostile environment from another colleague in CAIRF. The hubris is almost Shakespearean.
This is the man DSIT chose to present to the world as the guardian of ethical AI.
When I formally escalated concerns on 6th Februray 2025, DSIT’s Director General for Digital Technologies and Telecoms Emran Mian was made fully aware of Dean’s misconduct and hostility towards me. Emran had every opportunity to intervene, to hold CAIRF leadership to a higher standard. To be a leader himself. Instead, he insisted on hiding behind “culture review” nonsense.
Emran chose to shield and elevate Dean. A man who not only created and presided over a hostile environment, but who openly disrespected Emran himself.
On several occasions, Dean repeatedly confused and conflated Emran’s name with that of another senior official of South Asian descent. This wasn’t an innocent mistake. It was dismissive and disrespectful.
If Dean thought he was above learning the names of his direct seniors, two entirely different individuals, what does that tell you about him? His professionalism? His basic respect for the people he supposedly serves? If this is how he treats his seniors, how do you think he treats his juniors? Dean’s indifference wasn’t just tolerated. It was rewarded.
I stood up for my dignity. I stood up for the mission and integrity of CAIRF. For the credibility of the government’s standing as a “world-leader” in AI. I stood up for Emran himself in my initial email to him when I brought attention to the disrespect to his name happening behind his back. And for that, Emran chose to punish me. He chose to authorise retaliatory debts in my name. He chose to lock me out of my work mobile phone. He chose to surveil my email metadata when I was just trying (and failed) to get legal support. He chose to try to intimidate me with calls out of GCHQ Bude. He chose to treat me like a criminal.
What does all of this tell you about Emran’s judgement? His leadership capability? His character? His own self-respect? I’m not here to tell you what to think. I am simply delivering the facts and the truth. You be the judge.
Emran, here’s what you need to understand. Throughout all of this, I wanted to protect you. As I said in my first piece, I had so much respect for you and who I thought you were. Who you could’ve been. Those weren’t empty words. I really meant it. It’s sad you couldn’t also protect and respect me.
Like so many organisations, DSIT congratulates itself for encouraging its staff to include pronouns in their email signatures as a marker of ‘inclusion’, whilst real human dignity was left to rot. I did not need a pronoun badge to prove my womanhood. I needed a system that protected my dignity when it was being eroded. Whilst lives are being destabilised, whilst job markets face AI-driven collapse, DSIT thinks that performative gestures translate to progress.
If this is how DSIT treats a human employee raising legitimate concerns, if this is who they handpick to elevate, how can they be trusted to steward AI safety, impacting millions of people’s lives?
This is not just my story. This is a warning. A signal. About how the machinery of AI will inherit the machinery of human cruelty. About how the way AI governance is being shaped today mirrors the worst patterns of human cruelty, denial and institutional cowardice. The shadow.
They erased me from the Teams groups. But they cannot erase the truth.
You cannot run from your shadow forever.
I’m just getting started. See you tomorrow.
Follow me:
X/Twitter: @syro_001
Bluesky: @syro001.bsky.social
Mastodon: @syro001