John McDermid
@McDofYork
Followers
494
Following
222
Media
128
Statuses
1K
Professor of Software Engineering at the University of York. Research interests in safety, security and software engineering.
York
Joined January 2011
Now try that for addictive reasoning
0
0
1
Extraordinary. I wonder if “cracked” is the keyword here - or does it mean something different in American English?
Has there ever been a lower-energy web site than this? That's it. The whole thing. I can't tell if this is performative not-caring or genuine contempt for the public. https://t.co/Wszs8Jhqo8
0
0
0
Communication around assurance is often overlooked so this is a very valuable contribution to the topic
For #AIsafety, getting the basics right is essential. Here we reflect on the communicative nature of assurance: fundamentally a dialogue between individuals working together to achieve a shared understanding of risk, uncertainty, ... @CfAA_York
0
0
3
Many of the members of the HoL have done real jobs before joining the House and have knowledge and understanding to bring to bear as this shows
The House of Lords is a funny old place, but sometimes the people there say very sensible things. This is from one of the recent debates on the Automated Vehicles Bill https://t.co/PmHVpPrhKm
0
0
0
Bashar I think this is too delicate and subtle for X - but happy to discuss with you more directly
@McDofYork Who writes the safety requirements for civilians who ‘interact’ with this system, or are they not “stakeholders”?
1
0
1
Not sure about AI safety, but it wouldn’t pass ethical tests I would want to apply
@jmbromley @McDofYork @BarryOSullivan @DameWendyDBE @Abebab Hmm.. sounds like guns don’t kill people, people kill people…
2
0
1
Why cyclist crashed - they were stationary - and the car turning was on the wrong side of the road.
Officers from the Roads And Transport Squad are appealing for information following a Road Traffic Accident where a pedal cyclist crashed into a car that was completing a legal right turn https://t.co/DOSNtyGmZt
0
0
0
Interesting views - so still work for us to do on maritime assurance
0
0
1
Starting our Maritime launch event for the Centre for Assuring Autonomy
0
1
7
Yes, one of the highlights
Great to see @SteadyBits presenting our work on #ResponsibleAI and #AssuringAutonomy at the #TASShowcase2024. It is part of our @tas_hub project #TAS_RAILS with @Marenka @Carolyn_Ten @rdhawkins @Jackstilgoe and Jo-Ann Pattinson (@ITSLeeds)
0
0
1
Very worthwhile. Will results from this be published?
It is vital that responsible #AI policy incorporates the views of the wider public. At today's first People's AI Stewardship Summit, Academy staff are facilitating a workshop with AI stakeholders and members of the public to gather together and discuss the state and future of AI.
1
0
1
We’ve also done work on using LLMs to derive safety analyses, eg fault trees. There too more needs to be done but surprisingly promising
Some musings about - How Can Large Language Models Do a Fram Analysis? It's good fun to play around with these tools but lots more work required @SSRN
https://t.co/JiLhNDkCIF
@IHabli @AAIP_York @McDofYork
0
0
1
Look forward to listening
Out next month, @ensemblemoliere's debut album features works by Charpentier, Couperin, Lully & Marais. The King's Playlist is a luxuriant & varied selection of French Baroque music to accompany the Sun King’s day. You can listen to the first single now ► https://t.co/xmZpq6KvDR
0
0
1
Interesting. I’ve used causal loops on several occasions- always found them helpful to get a broad understanding of system dynamics
Been shared before, but really nice illustration of causal loop diagrams to help explain and visualise the negative impact of person-focused interventions https://t.co/vcEerGdDfm via @StevenShorrock
0
0
0
Is it a safe distance?
I reckon this is as close as I'll get to Jeremy Clarkson. I'm in @autocar
0
0
1