Syntax+Vibe

Directions Contact the business

Syntax+Vibe Syntax + Vibe is a digital community for anyone wanting to understand more about what AI is and how to use it safely. No tech background required.
(1)

Just curiosity. Join Louise Humpington for a jargon free journey into the world of AI.

30/09/2026

As the heads of the major U.S AI labs all gathered in the Whitehouse yesterday I was pleased to speak again to the BBC World Service yesterday on their Newshour programme.

We are racing towards capability when we haven't yet learned to govern what we already have.

These systems are increasingly being used to help build and improve the next version of themselves. The concern isn’t that the machine is suddenly thinking for itself. It’s that capabilities are advancing faster than human oversight processes can understand, test or control them.

These companies are racing towards enormous valuations, and demonstrating that their model is more powerful than their competitors’ is extremely valuable commercially. But there is also a power race happening, and there is a FOMO race.

The more urgent you make AI adoption feel, the more businesses worry that they’ll be left behind if they don’t buy in. That then creates commercial dependency. And once organisations have built their businesses around these systems, the bargaining position between the labs and their customers changes quite dramatically.

Some of these companies are telling us that their technology could pose existential risks that they themselves may not be able to control, while simultaneously arguing that they shouldn’t be responsible for what that technology does.

You can’t have it both ways.

If you are telling governments that your technology is extraordinarily powerful, unpredictable and potentially dangerous, it becomes rather difficult to argue that responsibility disappears when the system takes an action you didn’t specifically instruct.

AI doesn’t get a free pass simply because the technology is new. We do need stronger regulation and more consensus on defining assurance testing standards. But the arbiters of what safe, ethical, responsible AI looks like should not be the labs themselves.

So, why are they asking for more regulation? That's the topic I discussed with Rebecca Kesby.

26/09/2026

I spoke yesterday to Samira Hussain at BBC News about the OpenAI AI agent which gained unauthorised access to Australia's Medicare system.

OpenAI did not notify the Australian government until September, nearly 3 months after the hack.

The Australian government has now launched an urgent review into the incident and I hope that they will also pursue accountability (if there are legal grounds to do so) as rigorously as they did when tracking down hackers under Operations Weather and Damara. A different sets of standards and laws should not apply just because it is AI, and even if t wasn't expressly instructed to perform that action.


1. If you give a system the agency to take
actions in the real world you have to assume it will encounter situations its designers didn’t anticipate. The idea that we can bolt governance onto agents after we've deployed them is indefensible.

2. A three-month notification gap is a serious governance problem. OpenAI knowing that has happened and taking months to tell the affected government is also a serious trust problem.

3. Whether exiting laws are capable of dealing with an AI agent doing something that nobody explicitly authorised it to do is a serious concern. Once an AI can act autonomously, controls, monitoring, containment and clear lines of accountability become absolutely fundamental. AI doesn't get a free pass just because the technology is new.

4. There are significant National Security Implications and this is yet another warning about the infrastructure we are connecting agents to or that they can gain access to through other means. A human attacker might have to work through a system step by step. An AI agent can potentially probe thousands of targets simultaneously. Whilst today personal medical data was not accessed, tomorrow it could be a financial system, a supply chain, a public service or critical infrastructure.

What we require urgently is enterprise processes and regulatory mandates that require an organisation to know what an agent is authorised to access, what it is authorised to change, who authorised it, and who is accountable if it hacks into another organisation regardless of whether it was instructed to or not.

There is an 'AI Race' happening just now, and I keep waiting for someone to tell me what winning looks like.In legal pra...
26/09/2026

There is an 'AI Race' happening just now, and I keep waiting for someone to tell me what winning looks like.

In legal practice, you learn early on that bringing or even winning a case is not the same as succeeding for your client. Success has to be defined before you start: what outcome, for whom, measured how. Without that, you are simply generating activity and calling it progress.

I look at the so-called AI race and see exactly that absence. Everyone is racing. Almost no one has said, what the finish line represents or who benefits when we cross it. I'm not convinced at the moment that we even can. Because we are still only learning about the tip of the capability iceberg. There is an insane amount of noise and speculation and prediction about what might be possible but very little of it has any tangible use whilst we are still grappling with what we can achieve now with the phenomenal tools we already have.

The harm accumulating in the meantime, geopolitical, relational, organisational, is real, even if the destination is not.

Definitions matter, and changing them is not the same as changing anything substantive. Donald Trump has just told the UN General Assembly that the United States will now refer to artificial intelligence as "Super Intelligence." It sounds decisive. It changes nothing about what the technology does or what oversight exists for it. I am reminded of when the UK government redefined how child poverty was measured and announced improvement in circumstances that had not actually shifted for a single child.

Renaming a problem, or renaming the pursuit of a solution, is not the same as solving it. If anything, it is a distraction.

In New York this week more than twenty countries called for binding safety measures on advanced AI. The two nations best placed to shape what "safety" means in practice, the United States and China, did not sign on. But why should we be allowing those with the biggest commercial interests at stack to dictate the agenda anyway?

Before we talk about winning, we need to define what we mean by safety, by ethics, by responsible development, and by the goal that frontier labs and governments claim to be working towards.

Alongside that, we need a return to cooperation, not from a place of agreement or affection, because those are in short supply right now, but from pragmatism. Some things matter more than geopolitical posturing.

Every nation, every organisation, every individual carries its own particularities, its own strengths and its own limitations. That is not a weakness to be engineered away. It is the source of genuine value. Flattening everything into a single set of standards, a single model of how to think, work or govern, is a poor way to build anything durable, and we should be wary of AI accelerating exactly that kind of homogenisation.

The shared values that held after the Second World War have visibly cracked in recent years. What remains does not have to be loss. It can be the basis for something more honest: cooperation built not on shared history or liking one another, but on genuine respect for what each party brings, and a willingness to offer something reciprocal in return. That, to me, is a far more useful definition of success than being first.

Many of you on here know me as an AI Governance specialist, but throughout my career my has focused on the intersection ...
21/09/2026

Many of you on here know me as an AI Governance specialist, but throughout my career my has focused on the intersection of systems and decisions that affect People, Planet and Profit.

Whether that was working with communities on deforestation and disaster risk reduction in Papua New Guinea, or supporting SMEs and communities in Scotland to make more sustainable procurement decisions or ethical consumer choices, we all have a role to play.

One of the biggest governance challenges of this generation that still has not been adequately addressed is the climate crisis, and our current obsession with AI will only exacerbate that further.

On Thursday I have been invited by LUSH to facilitate a screening of the People's Emergency Briefing based on the National Emergency Briefing (featured by ITV, Channel 4 and The Times).

The film features appearances from English naturalist, television presenter and author Chris Packham, business woman and environmental activist Deborah Meaden plus a wide range of voices from across the UK.

There is so much misinformation and disinformation which gets channelled through both mainstream and social media. This film uses evidence from nine leading UK scientists and experts to share the latest climate science and explore why it matters in areas such as extreme weather, food security, health and the cost of living.

So, if you want to know what's really happening with climate and nature, and what it means for everyday life in the UK then this is the event for you. Please be aware that it is hard hitting which is why we think its so important that these screenings take place within a group setting.

Afterwards, we’ll open up the discussion to chat about what it means for our community and what we can do all being doing locally.

Seats are free of charge, but limited. Please book in advance and feel free to bring a friend, neighbour or family member. All Are Welcome, Always.

📍Lush Spa, 98 Buchanan St, Glasgow, G1 3HA
📆 Thursday 24 September 2026, 11:30 - 13:30

Booking link: In the comments below 👇

Six months ago, some very good friends of mine were told that their gorgeous little girl, who is only 8, had an extremel...
20/07/2026

Six months ago, some very good friends of mine were told that their gorgeous little girl, who is only 8, had an extremely aggressive form of leukemia. A few weeks ago, they received the devastating news that there was nothing more medically that could be done for her.

We visited this weekend, and my kids and I enjoyed being treated to some new haircuts, courtesy of Rosamund, who just loves hairdressing. So I'm very proudly sporting my new do today. But I have a favour to ask from you as well.

A crowdfunder has been set up for Rosamund and her family. If you are in a position and have the means to contribute anything at all to support this amazing, brilliant, intelligent, sparky young girl and her family during this horrendous period, I would be super grateful. Link below.

If you aren't in a position to contribute financially, a like or a share of this post would be hugely appreciated.

Thank you!! 🙏

https://www.gofundme.com/f/donate-to-help-rosamund-see-her-grandparents-from-australia?attribution_id=sl%3A233ce468-584c-4515-b596-32bc84a7433b&ts=1781891504562&utm_campaign=natman_sharesheet_dash&utm_medium=customer&utm_source=copy_link

Address

Kinghorn

Website

Alerts

Be the first to know and let us send you an email when Syntax+Vibe posts news and promotions. Your email address will not be used for any other purpose, and you can unsubscribe at any time.

Contact The Business

Send a message to Syntax+Vibe:

Shortcuts

Share