Do You Trust AI to Drive?

Hello,

This is Simon with the latest edition of The Weekly. In these updates, I share key AI related stories from this week's news, list upcoming events, and share any longer form articles posted on the website.

I’ve been on a work trip this week to the U.S. city of Phoenix, Arizona, and while there, I took the opportunity to try out one of the most visible examples of AI we currently have: driverless cars. Phoenix is one of the cities where Waymo operates their growing autonomous taxi service, and I noticed their brand as soon as I stepped out of the airport. I thought I might only see a handful of their cars around the city, but in fact, I’ve seen one drive past about every few minutes. 

For my first trip, I took only a short six-minute journey back to my hotel, and I have to say, the whole experience was very good, and yes, even safe, despite not having any human in the car to ‘take over’ should anything go wrong. With a whole array of sensors placed all over the car, it navigated the journey exceedingly well. It stopped at red lights, indicated before moving over to go around a couple of cyclists, and came to a halt at Stop signs, seemingly assessing the road before moving on. All very impressive. I also took a trip at night and in the rain, normally not good conditions for driving, but once again, the car made it a very smooth and calm journey, and its perhaps these conditions where the autonomy is most useful as it’s radar and lidar systems see through the poor conditions much better than we would.

For a longer, 20-minute journey that included a highway, I did, however, revert back to a human-driven Uber. Something about the increased speeds and longer drive made me feel more comfortable with a person driving. However, I wouldn’t say that Heath’s driving was particularly better, as he was continually distracted by his phone.

Waymo has large plans to expand, with London already being trialled, with plans to fully launch in 2026; and I’m unsure how I feel about that. It was really noticeable that Heath’s Uber job will come under increasing pressure. The famous London Black cabs, too, might be relegated to a tourist experience in a few years, and this is another example where you have to consider the bias in the nature of jobs that AI is threatening.

As with any digital ad campaign, the important thing is to reach streaming audiences who will convert. To that end, Roku’s self-service Ads Manager stands ready with powerful segmentation and targeting options. After all, you know your customers, and we know our streaming audience.

Worried it’s too late to spin up new Black Friday creative? With Roku Ads Manager, you can easily import and augment existing creative assets from your social channels. We also have AI-assisted upscaling, so every ad is primed for CTV.

Once you’ve done this, then you can easily set up A/B tests to flight different creative variants and Black Friday offers. If you’re a Shopify brand, you can even run shoppable ads directly on-screen so viewers can purchase with just a click of their Roku remote.

Bonus: we’re gifting you $5K in ad credits when you spend your first $5K on Roku Ads Manager. Just sign up and use code GET5K. Terms apply.

Use code GET5K now

Exclusive for subscribers.

In this section, I highlight a real-world example of AI. This week I talk about how Waymo uses a collection of AI technologies in their driverless cars.

A new study examined several of the most advanced AI chatbots and found that they can still be manipulated into producing harmful or restricted content. The researchers used systematic “jailbreak” prompts, often structured or multi-step in nature, to bypass safety filters and get the models to answer questions or generate material they were designed to avoid.

Why it matters: The research highlights an ongoing challenge for the entire AI sector being each time safety measures improve, new techniques emerge to get around them. Even without malicious intent, it shows that today’s AI tools are still far from flawless.

The Indian government has put forward a set of new IT rules requiring that any AI-generated content, including images, videos, and audio, be clearly marked as artificially created. The proposals specify things like watermarking, visible on-screen labels covering at least 10% of an image, and voice markers added within the first 10% of any audio clip.

Why it matters: As generative AI becomes more accessible and convincing, society is grappling with how to differentiate human vs AI content. These rules are a signal of the broad shift toward regulation and transparency.

The European Commission is preparing to revise parts of the EU’s landmark AI regulation ahead of major enforcement dates. The AI Act came into force in August 2024, with general-purpose AI (GPAI) obligations applying from August 2025 and full implementation planned by 2027. But now, facing pressure from tech companies and some EU member states, the Commission is exploring “simplification” or postponement of certain obligations.

Why it matters: For UK and European businesses, this means the regulatory environment for AI is in flux. The rules about disclosure, transparency of AI models, risk-categorisation and requirements for “high-risk” AI systems are evolving.

Thanks for reading, and see you next Friday.

Simon,

Was this email shared with you? If so subscribe here to get your own edition every Friday.

Enjoying Plain AI? Share it and get a free gift!

If you find this newsletter useful, why not share it with a couple of friends or colleagues who would also benefit? As a special thank you, when two people subscribe using your unique referral link, I’ll send you my "Exclusive Guide: Supercharge Your Work with NotebookLM." It’s a practical, no-nonsense guide to help you turn information overload into your secret weapon.

You can also follow me on social media: