AI Scams Are Getting Worse, and Healthcare AI Has a Big Blind Spot

If you’ve gotten a suspicious email lately that felt a little too polished, you’re not imagining things. We’re firmly in the new era of AI-driven scams, and it’s not pretty.

When ChatGPT dropped in late 2022, it didn’t take long for cybercriminals to realize they could use it to write phishing emails that didn’t read like they were composed by someone who failed middle school English. Since then, they’ve expanded to turbocharged phishing, hyperrealistic deepfakes, and even automated vulnerability scans. Basically, if there’s a way to abuse AI for profit, someone’s already doing it.

The volume of attacks is overwhelming many organizations. AI makes everything faster, cheaper, and easier to scale. And this problem is only going to get worse as more criminals adopt these tools and the underlying models improve. Rhiannon Williams has the full story over at MIT Technology Review, and it’s worth a read if you want to understand just how fast this is escalating.

This “supercharged scams” trend is one of the 10 Things That Matter in AI Right Now, their essential guide to what’s actually worth paying attention to in the field. Subscribers can also watch a roundtable discussion with their AI reporter and executive editors.

Meanwhile, on the healthcare side of AI, there’s a different kind of problem. Doctors are using AI for notetaking, combing through patient records, flagging people who might need specific treatments, and interpreting medical exam results and X-rays. A growing number of studies show these tools can deliver accurate results. But here’s the uncomfortable question nobody seems to want to answer directly: Does using them actually translate into better health outcomes for patients?

Spoiler: we don’t have a good answer yet. Jessica Hamzelou explains why in her piece for The Checkup newsletter. It’s a frustrating situation — we’re deploying these tools at scale without solid evidence that they’re making people healthier. That’s a big blind spot for an industry that’s supposed to be all about evidence-based practice.

And in other tech news today: DeepSeek finally unveiled its long-awaited new AI model. Preview versions of DeepSeek-V4 are out, and the company claims it’s the most powerful open-source platform, rivaling top closed-source models from OpenAI and DeepMind. It’s also adapted for Huawei chip technology, which is an interesting move given the ongoing chip restrictions. Bloomberg and SCMP have the details if you want to dig deeper.

More countries are moving to curb children’s social media access. Norway is set to enforce the latest ban, and the Philippines could follow soon. Meanwhile, Americans are pushing to get AI out of schools. The backlash against tech in education is real and gaining momentum.

That’s the landscape today: AI making scams worse, healthcare AI running on faith rather than evidence, and governments scrambling to protect kids from the platforms they helped create. Not the most cheerful roundup, but at least it’s honest.