OpenAI Cybersecurity AI, ASML Forecast & Samsung AI TVs

Samantha: Welcome to The State of Tech, The European Edition, Thursday April sixteen, 2026. I'm Samantha Lawrence.

Bob: And I'm Bob Russell. Today: OpenAI launches a specialized AI model built for cybersecurity, ASML raises its sales forecast on surging AI demand, one in four Americans now ask AI for health advice, Volcengine pushes the concept of an AI Knowledge Brain for enterprise, Microsoft hardens Windows against phishing attacks through Remote Desktop, and Samsung shows off AI-powered TVs in Frankfurt that want to become your living room companion. Let's start with cybersecurity.

OpenAI enters the cybersecurity arena with a dedicated AI model

Samantha: So OpenAI has just unveiled something called GPT-5.4 Cyber. And this isn't a general-purpose chatbot. This is a model specifically built for cybersecurity professionals. We're talking vulnerability detection, malware analysis, hardening digital defenses. It's being rolled out under what they call the Trusted Access for Cyber program, or TAC, and right now it's limited to vetted security experts.

Bob: This is a really significant shift in how we think about AI in security. Up until now, most of the conversation has been about AI as a threat — deepfakes, automated phishing, AI-generated malware. And those are real problems. But now OpenAI is essentially saying, okay, let's put this same power on the defensive side. Give the good guys a serious tool to fight back. The model can apparently analyze code for vulnerabilities at a speed and depth that would take a human team weeks. That changes the economics of cybersecurity.

Samantha: And this is where it gets interesting for Europe. European critical infrastructure — energy grids, financial systems, hospitals — faces constant cyberattacks. A tool like this could be a game-changer for understaffed security teams. But there's a tension here. The EU AI Act is rolling into enforcement, and a powerful AI model operating in sensitive security contexts is going to raise flags. Who's accountable when the AI misses a vulnerability? Or worse, when threat actors find a way to manipulate it?

Bob: There's also the access question. Right now it's limited release, which means OpenAI is controlling who gets in. European governments and agencies will want to know — are they getting equal access? Or is this another case where American entities get first dibs? For European cybersecurity firms, there's also a competitive angle. If OpenAI dominates this space, where does that leave homegrown European security AI?

Samantha: And let's not forget the data side. If European organizations feed sensitive security data into an American AI model, that raises GDPR questions immediately. Where is that data processed? Who can access it? These aren't hypothetical concerns. They're the kind of thing that could slow down adoption even if the technology is brilliant.

Bob: The bottom line is that the AI arms race in cybersecurity just escalated. Defenders now have a serious new weapon, but deploying it responsibly — especially in Europe's regulatory environment — is going to require careful thinking.

---

ASML lifts its outlook as AI spending shows no signs of slowing

Bob: Next up — ASML, Europe's crown jewel in semiconductor technology, just raised its full-year sales forecast. They now expect net sales between 36 and 40 billion euros, up from a previous range of 34 to 39 billion. The reason? AI demand that just keeps accelerating.

Samantha: This is one of those stories that really shows you how deep the AI boom goes. ASML doesn't make chips. They make the machines that make the chips. Without ASML's extreme ultraviolet lithography systems, you simply cannot produce the most advanced semiconductors. Period. So when ASML raises its forecast, it's telling you something fundamental about where the entire industry is heading.

Bob: And these numbers are not trivial. We're talking about potentially 40 billion euros in revenue for a single company that most consumers have never heard of. Every major chipmaker — TSMC, Samsung, Intel — depends on ASML's equipment. So this forecast revision is essentially a vote of confidence that AI infrastructure spending will remain intense through the rest of 2026.

Samantha: For Europe, this is a huge deal. We often hear criticism that Europe missed the boat on tech — no European Google, no European Apple. But ASML is proof that Europe sits at an absolutely critical chokepoint in the global tech supply chain. Without Veldhoven in the Netherlands, the AI revolution stalls.

Bob: There's a geopolitical dimension too. ASML is already subject to export restrictions to China, and as AI becomes more strategically important, the pressure on ASML — and on the Dutch and EU governments — is only going to increase. Being indispensable is powerful, but it also paints a target on your back.

Samantha: Still, for European investors and policymakers, this is encouraging. It shows that Europe can compete at the highest level of technology when it plays to its strengths.

---

Samantha: Quick interruption. If you listen to The State of Tech regularly, hit that like button and subscribe — that way you'll never miss an episode. Okay, moving on.

A quarter of American adults now turn to AI chatbots for health advice

Samantha: Here's a number that stopped me in my tracks. Over 66 million American adults — that's one in four — are now using AI tools for health information. Physical health, mental health, you name it. This comes from new research by the West Health-Gallup Center.

Bob: And what's interesting is how people are using it. More than half of those 66 million say they use AI to supplement their doctor visits, not replace them. They'll go to a chatbot before an appointment to understand symptoms, or after an appointment to make sense of what the doctor told them. It's almost like a health translator.

Samantha: Some of the motivations are pretty telling though. Quick answers and curiosity are top reasons, sure. But a meaningful number of people said they turn to AI because of cost or access barriers. They can't afford a doctor visit, or they can't get an appointment fast enough. That's a commentary on the American healthcare system as much as it is on AI adoption.

Bob: Now, for Europe this is a preview of what's coming. European healthcare systems are different — most have universal coverage — so the cost motivation may be less dominant. But access? Waiting times? Those are very real problems across Europe. If you're waiting six weeks for a GP appointment in parts of the UK or the Netherlands, the temptation to ask an AI chatbot is obvious.

Samantha: The risk is equally obvious. AI chatbots can sound incredibly confident while being completely wrong about a medical condition. And the consequences of bad health advice are not like getting a bad restaurant recommendation. European regulators will need to think carefully about how to handle this — possibly classifying health-focused AI as high-risk under the AI Act.

Bob: It also raises an interesting question about trust. If millions of people trust AI for health advice, what does that mean for the doctor-patient relationship? This is a societal shift, not just a tech trend.

---

Volcengine pushes enterprises to build an AI Knowledge Brain

Bob: Let's talk about enterprise AI strategy. Volcengine — that's ByteDance's cloud and AI division — is advocating a pretty compelling framework. Their VP Yang Liwei says companies need to build what he calls an AI Knowledge Brain and an AI Execution Manual. The idea is to consolidate all your business rules, system interfaces, and institutional knowledge into a central hub that AI can actually use to get things done.

Samantha: So rather than bolting AI onto existing processes, which is what most companies do, you're essentially rebuilding the operational logic of the business around AI as the primary actor. The AI doesn't just advise — it executes through APIs, and the results feed back into the system. It's a closed loop.

Bob: Yang also teased an AI-centric human-machine interaction solution for vehicles, which is interesting because it signals Volcengine wants to go beyond enterprise software into physical products. Proactive, automated interactions in EVs — your car anticipating what you need before you ask.

Samantha: For European businesses, this framework is worth paying attention to. A lot of European enterprises are struggling with AI adoption because they're trying to layer AI on top of legacy systems. The Knowledge Brain concept suggests a more radical approach — centralize first, then automate. That could resonate especially with large German manufacturers or French logistics companies sitting on massive amounts of institutional knowledge that's currently locked in documents and people's heads.

Bob: And the automotive angle matters hugely for Europe. If Chinese EV makers start shipping cars with deeply integrated AI assistants that learn and anticipate, European automakers need to match that or risk falling further behind in the software-defined vehicle race.

Samantha: Competition from ByteDance's ecosystem is something European companies can't afford to ignore.

---

Microsoft strengthens Remote Desktop security to fight phishing

Samantha: Also today — Microsoft has rolled out enhanced security protections for Remote Desktop Protocol files in Windows. This came with the April 2026 security update, and it targets a specific attack vector that hackers have been exploiting: malicious RDP files sent through phishing emails.

Bob: For anyone who's not familiar, RDP is how millions of people connect to remote computers — it's essential for remote work, IT administration, and cloud access. The problem is that attackers have been crafting malicious RDP files that, when opened, silently share your local drives, clipboard, and other resources with the attacker's machine. You click one file, and suddenly someone has access to everything on your computer.

Samantha: The fix is actually pretty straightforward. Microsoft now shows much stronger warning dialogues before an RDP connection is established. You see detailed information about what the remote system is requesting, and you have to explicitly approve any resource sharing. It's friction, but it's the kind of friction that saves people.

Bob: For Europe, this is important on a practical level. RDP is everywhere in European businesses — from small accounting firms to major banks. Every organization that uses remote access benefits from this. And it aligns nicely with the broader European emphasis on cybersecurity. ENISA, the EU's cybersecurity agency, has been flagging remote access vulnerabilities for years.

Samantha: It's also a reminder that sometimes the most impactful security improvements aren't glamorous. No AI, no blockchain. Just a better warning dialogue that makes people stop and think before clicking. Sometimes that's enough.

Bob: Simple, effective, and long overdue. That's the kind of security update we like to see.

---

Samsung's AI-powered TVs want to be your new housemate

Bob: And finally, something a little lighter. Samsung has been showing off its 2026 TV lineup at a tech seminar in Frankfurt, and these aren't just televisions anymore. They're... companions? Samsung is calling the new platform Vision AI Companion, or VAC, and it basically turns your TV into an AI assistant that watches along with you.

Samantha: Okay, I have to admit this is both cool and slightly unsettling. The TV can provide real-time information about what you're watching — like identifying an actor or explaining a scene — and it can help with everyday activities. Samsung says it makes the screen more adaptive and responsive. So your TV is essentially paying attention to your life.

Bob: I mean, we've gone from arguing about who has the remote to the TV deciding what we should watch next. Progress, right?

Samantha: Ha! I'm imagining the TV judging my late-night snack choices. "Samantha, based on your viewing pattern and the time, might I suggest a salad instead?"

Bob: That said, the technology is genuinely impressive. Samsung is rolling AI capabilities across its entire 2026 lineup, not just premium models. That democratization matters. And hosting the seminar in Frankfurt shows how seriously Samsung takes the European market.

Samantha: The European angle here is interesting though. European consumers tend to be more privacy-conscious than average. A TV that's constantly analyzing what you're watching and doing? That's going to spark conversations about data collection. How much does it store? Where does it go? Under GDPR, Samsung will need to be very transparent about this.

Bob: True. But if they get the privacy balance right, these could be genuinely useful. Imagine an elderly person whose TV can detect a fall or remind them to take medication. There's real potential beyond entertainment.

Samantha: As long as it doesn't start critiquing my taste in reality shows, I'm cautiously on board.

---

Samantha: That's our show for today. We covered OpenAI's new cybersecurity AI model, ASML's raised forecast on AI demand, Americans turning to chatbots for health advice, Volcengine's AI Knowledge Brain framework, Microsoft's RDP security upgrade, and Samsung's AI companion TVs debuting in Frankfurt.

Bob: Want to know more or react? Visit stateoftech.eu or email us at info@doorzetters.net.

Bob: State of Tech — the tech world in 15 minutes.