Some of the biggest names in AI are now asking for regulation. Sam Altman and Dario Amodei are calling for the industry to slow down, while Jensen Huang at Nvidia says there is no problem and we should speed up. So what is really going on, and what does it mean for how MSPs, IT teams and business owners defend themselves?
In our latest RoboShadow video, Terry Lewis walks through the debate, responds to investor Steve Eisman's view on the commercial interests behind AI regulation, and explains the practical changes he thinks every security stack needs to make now.
Terry covers new topics like this every week, so if you find this useful, subscribe to RoboShadow on YouTube to catch the next one.
Terry is clear that there is real danger in AI, but he thinks much of the conversation is blown out of proportion. When the leaders of the main frontier model companies say AI needs to be reined in, and the CEO of the company supplying the chips says the opposite, something else is clearly going on in the middle.
Steve Eisman, the well-known investor, argues that the "Terminator" narrative is a distraction. In his view, open weight models are taking market share, the big AI companies have no real moats around their business, and they may be trying to create regulation they can then shape to protect their position.
Terry broadly agrees. Chinese AI models are very good. They are not quite at the level of the frontier models, but they track them closely, and because AI is now training AI, they do not need to copy the frontier models to keep improving. As long as they have the compute, they can keep self-training. That gives the frontier companies a commercial reason to want Western regulation that limits which models can be used.
Jensen Huang, meanwhile, wants to go full speed, and with good reason: the Chinese models that worry the frontier labs run on his technology too, so Nvidia wins either way.
Even so, Terry believes regulation is genuinely needed. The commercial motives on every side do not change the fact that more AI will get out of its sandbox and more organisations will be hit.
For anyone responsible for protecting a business, Terry's message is that logic needs to prevail. Your core cyber security fundamentals are still very important, but it is time to take them one stage further. This is why RoboShadow is building three new capabilities.
Want to see how your own fundamentals look today? In a couple of minutes you can get your vulnerabilities, devices and Microsoft 365 posture in one place.
Create your free RoboShadow account and check your fundamentals ➡️
AI is not a ghost. An AI-driven attack still has to enumerate systems and pull data from APIs, and that leaves a signature. You need to detect that signature and shut it down quickly. SIEM and SOC capabilities are now going live in RoboShadow to do exactly that.
Our AI Pentest Simulation chains your live vulnerabilities into the attack paths an attacker could take, from external recon right through to domain admin, so you can see exactly where to close the gaps first.
Every business line is picking up new tools, from NPM packages to small, esoteric applications and shadow AI services. With RoboShadow App Protect, you will be able to see which applications are being used across the organisation and shut down the ones that should not be there. Highly security-conscious teams can run in deny mode first and then build an allow list, while others can use the visibility to decide which parts of the business can use which technology, keeping shadow IT to a minimum.
Would you like an early look at App Protect and our AI SIEM and SOC work? We will walk you through where the roadmap is heading.
Terry's conclusion is a sensible one. This is not the end of the world and it is not a reason to panic. The people profiting from AI have their own interests in how regulation plays out, and it pays to look past that. What matters is that AI threats will keep growing, so your defences need to grow with them. If you have strong fundamentals in place, can detect AI-driven activity, test yourself with AI, and control what runs on your devices, you will be in a good position.
RoboShadow is a cyber security platform that helps businesses and MSPs identify, prioritise and remediate vulnerabilities across their IT environments. It combines internal and external vulnerability scanning, device security, Microsoft 365 integrations and automated remediation, helping teams manage cyber security continuously while reducing cost, complexity and manual work.
Adjusting your stack for AI does not have to be a big project. Here are three easy ways to take the next step today:
As always, thank you for your ongoing support and feedback. If you have any questions, please don't hesitate to reach out to us at hello@roboshadow.com or get in touch with our team. We are always more than happy to walk through things with you.
Watch more from RoboShadow on our YouTube channel.