Just reading about how after Agincourt, when Henry V returned there were fountains of wine in the streets of London. You have to say, maybe that would be a great way to silence all of the doom sayers.
Nobody, and I mean nobody talks themselves down quite like the British. One day we might discover positivity and optimism, but I doubt it.
Caution with appropriate safeguards and governance clearly the best approach here. GenAI will be used regardless, and it will take time to address sufficiently
Just absolutely buzzing for this. Don’t care what the negative commentary says. Legend.
Data being exposed to the web is immeasurably harmful. No doubt the terms for the software exclude almost all liability too.
Maybe the juice isn’t worth the squeeze sometimes?
cybernews.com/security/emp...?
An employee monitoring service has been leaking millions of screenshots as part of a rather large data breach.
As an in-house lawyer, this would keep me awake much more than whatever the individual employees are doing on a day to day basis.
All of your confidential and commercially sensitive
I really like the woman who is screaming in my header graphic here.
AI governance from scratch, the outline of getting something up and running.
open.substack.com/pub/neverher...
I have to say, I absolutely rate the new Devil May Cry series. I loved the games and I bloody loved that.
Special mention to the white rabbit.
All to play for in the second-half at Villa Park. ⏸️
My belated thoughts on the Waymo saga that happened ages ago (last week).
open.substack.com/pub/neverher...
I think there should be more of a focus on transparency in AI. I don’t just mean a banner to tell people you’re working with it, though they’re good too. I’ve popped my thoughts below:
neverhere.substack.com/p/transparen...
Apparently the AI Act could be up for amendment when most of it is not even applicable yet.
www.politico.eu/article/how-...
Less problematic and contentious.
Unless of course it is designed for the more difficult use cases and they did intend to implement predictive policing. Then disregard the above and we are in minority report.
www.theguardian.com/uk-news/2025...
Deployment and development of these kinds of systems and use in the public sector.
If there had been a release which explained the full intended use of the system, the mitigation measures put in place (presumably human review) and proactive ongoing monitoring, I suspect it would be…
An interesting article, which Yvette Cooper was asked about this morning.
Using AI to allocate resources can still be problematic of course, but it is much less concerning than some of the potential uses suggested by the article.
It just highlights the importance of transparency in the…
Got to say out of the Studio Ghibli saga one of the things that stands out to me is what trade secrets/confidential information has been inadvertently hoovered up OR just input into an AI system by company employees.
Amazon’s GC was right to be concerned it seems.
Thought I was about to read an interesting take on how a law firm was using AI Agents, but it turned out to read like AI generated slop with almost no depth whatsoever.
Nice to know google alerts are still 50/50.
www.londondaily.news/how-ai-agent...
A law in New Jersey has been signed which establishes criminal and civil penalties for those who produce AI Deepfakes. Largely this is in response to fake pornographic images but also captures images made for political manipulation.
6abc.com/post/ai-deep...
The problem with the 23andme saga is that everyone said this was a massive risk and nobody cared and did it anyway.
Will be interesting to see if these kinds of explanations can lead to solutions with strange outputs/personal data fabrications. Could prevent issues such as the recent ChatGPT case re hallucinations and defamation.
The more I read about AI agents the more I am interested in the legal ramifications of it all.
Is my AI Agent going to be working on my behalf or the company who built it?
What if during its task execution something comes up which conflicts between the two? Who bears that risk?
Companies are reportedly looking to move away from US based hyperscalers and cloud companies due to security concerns and the US Administration.
www.wired.com/story/trump-...
Any data used is non-biased.
There must be clear and unambiguous health warnings when interacting with AI.
The Act would create an AI Authority, similar in nature to the ICO presumably.
A company must supply to the AI Authority a copy of all its training data and IP used in the training of it.
Lord Holmes private members bill on AI Regulation has been reintroduced. While very unlikely to become law, it may shape debate.
Interesting points include:
Appointment of AI responsible officers if your company develops, deploys or uses AI systems to ensure non-discriminatory use of AI and….
Nvidia have set out that any videos generated by Cosmos world models will include digital watermarks.
These are in efforts to preserve IP rights and also to fight against misinformation.
www.theverge.com/news/631547/...
The Virginia Legislature have passed a bill restricting minor use of social media to one hour a day.
I wonder whether some of the criticism about the over regulation of the EU will calm down a touch in light of the above.
Looking forward to seeing the anti-health and safety brigade praising Sadiq here.
Any minute now I’m sure.
