Feed aggregator

• Dali's new Kupid stereo speakers cost £299 (about $400 / AU$625)
• 4.5-inch woofer with 26mm custom-designed tweeter
• 4-ohm impedence makes them easy to play from any stereo amp

Danish hi-fi company Dali has just announced a new pair of budget stereo bookshelf speakers, and as someone who's been testing the model just above them in the range at home recently, I think these could be very interesting.

The new speakers are called the Dali Kupid, and they're hoping to make you fall in love with their funky color options (as well as more traditional wood finishes), and their promise of detailed, audiophile-pleasing sound for a low price of just £299 (about $400 / AU$625 – no price has been confirmed for the US or Australia at the time of writing).

They sit between the cheapest Dali Spektor 1 bookshelf speakers ($280 / £199 / AU$499) and the impressive Dali Oberon 1 speakers ($600 / £399 / AU$749) – both of which scored five stars from our friends at What Hi-Fi? in their reviews of those products.

I've been using the Dali Oberon 1 at home recently as part of testing a new streaming amp, and they're really quite astounding for the price when it comes to detail and musicality – which means I think these genuinely could be fantastic value.

The Kupid are built with a custom-designed 26mm tweeter paired with a 4.5-inch mid-woofer. While the woofer appears to be very similar to one in the Spektor 1, the combination with a new tweeter and different bass reflex design could produce wider-ranging sound than the Spektor 1.

Though having said that, I should note that the Kupid are rated for slightly less extensive bass frequencies than the Spektor 1 (63Hz for the Kupid and 59Hz for the Spektor) – but spec numbers never tell the whole story with speakers, so I would expect a more full sound from the Kupid when factoring in all elements of the design.

One of the big focuses of the Kupid seems to be making them easy to live with – Dali suggests they should be pretty unfussy to place and get good sound from, and they come with wall brackets in the box as well as rubber feet.

They're also reasonably small, and they can be powered comfortably from 4-ohm amplification, so budget amps should have no problem getting their best sound. Dali says they should sound great quiet as well as loud, so they're suited to a lot of different environments – this is something the Oberon 1 are great at, so I don't doubt it here.

And perhaps coolest of all, they come in five great finishes for different tastes: Black Ash, Walnut, Caramel White, Golden Yellow, and Chilly Blue.

(Image credit: Dali)A great way to step up to the detail of bookshelf speakers?

These look like they would pair nicely with something like the Pro-Ject Stereo Box E amp, which costs $349 / £199 and should have enough power for these speakers, plus has Bluetooth built in. That would get you lovely analog audio components from two great hi-fi makers, all in a compact size, plus the convenience of wireless connectivity – all for under £500 total, in the UK.

If you compare to a similar wireless stereo speaker setup – something like a pair of Sonos Era 100 speakers costs only a little less – you'd almost certainly hear a major difference in detail and clarity from the hi-fi system.

With far bigger speaker drivers, more air moved, and more space for powerful components, you'll find that music has a lot more room to express itself than from a smaller system. This usually means you'll get the experience of hearing 'new' elements in songs, or just be able to appreciate them anew with an improved sound profile.

I actually did a comparison listening test recently between a pair of stereo Sonos Era 300 speakers and the Dali Oberon 1 speakers with the new Wiim Amp Ultra powering them – and although the Sonos speakers gave a great account of themselves with the sound dispersal and hefty low-end, the Oberon 1 had a clear edge when it came of mid-range expression, detail across the frequencies, and the handling of complex instrument mixes. Basically, all the things that make you feel really immersed in a song are boosted.

Obviously, we'll have to give these a real test to see if they can do the same at a lower price, but given my recent experience with Dali speakers, and the company's history, I think these look like a good threat to our list of the best stereo speakers.

You might also like…

• Dali’s mega-powerful 16-inch ported subwoofer wants to shake your home theater to bits, in a good way
• Sonus Faber Concertino G4 review: the bookshelf-speaker dragon I’ll be chasing for decades to come
• Pro-Ject’s new cheap turntable cuts out the bells and whistles to focus on sound – but it might be one cut too far

• 64-bit Google TV apps should deliver smoother, faster performance
• Developers told to prepare for "upcoming" Google TV/Android TV devices
• Target date is August 2026

Google has asked app developers to prepare for "upcoming 64-bit Google TV and Android TV devices" by making sure their TV apps are available as 64-bit versions.

Developers have plenty of time to prepare: the new rules come into force in August 2026.

Moving from 32-bit to 64-bit is good news for Google TV and Android TV users with compatible hardware, because 64-bit apps generally deliver faster loading times, less lag and better overall performance.

And streaming fans can see that in action on an Apple TV 4K, because Apple started the 64-bit app transition ten years ago and completed the switchover in 2019.

Many third-party devices, such as the Nvidia Shield, are 64-bit ready. (Image credit: Future)Does Google's 64-bit move mean new hardware?

Yes, but not necessarily from Google: while there's been some speculation that Google is working on a new 64-bit Google TV Streamer for launch next year, the operating systems are also used by third party products such as the Nvidia Shield and several of the best TVs, including the Sony Bravia 8 II. Google's blog post notes that three versions of the Nvidia Shield are 64-bit capable.

While Apple removed 32-bit support in tvOS 13 back in 2019, Google isn't following suit. "We’re not making any changes to 32-bit support, and Google Play will continue to deliver apps to 32-bit devices," Google TV product manager Fahad Durrani wrote.

What Google is doing here is asking developers to futureproof their apps, and from next August that means submitting both 32-bit and 64-bit versions for maximum compatibility.

That means users of existing Google TV and Android TV hardware don't need to worry about their apps disappearing or being left without updates in the foreseeable future, but depending on your device it might mean a performance boost is coming next year – or it might mean your next device gets the boost.

You might also like

• The Google TV Streamer is now finally worth buying in the UK
• The best streaming devices
• The best TV, chosen by our reviewers for all budgets

• Swann has launched a wireless home security camera with solar charging
• The EVO Wireless Solar can be powered by 45 minutes of sunlight per day
• The camera also has 16GB local storage, so you don't need a subscription

Swann, the company behind some of the best home security cameras we've tested here at TechRadar, has launched a new compact outdoor camera that you'll never need to take down for recharging, and which doesn't require a subscription plan to save and review your video footage.

The Swann EVO Wireless Solar is a compact weather-resistant camera (much smaller than the all-seeing Swann MaxRanger4K Solar), and can run all day with just 45 minutes of sunlight exposure, giving you plenty of flexibility as to where you mount it.

The camera comes with a 16GB SD card, so you can save your footage locally and keep full control over it. If you do decide that you want to keep clips in the cloud, the free Swann Secure plan gives you 1-7 days of cloud recording for a single camera.

The EVO Wireless Solar records at 2K with a 120-degree field of view, and offers infrared night vision for spotlight-free recording after dark. There's two-way audio as well, letting you speak to visitors and warn off potential intruders.

(Image credit: Swann)

The Swann EVO Wireless Solar has a list price of $129.99 / £99.99 / AU$179.95, and is available to buy direct from Swann.

For comparison, the Ring Outdoor Camera Plus Battery is $79.99 / £79.99 / AU$179, but lacks solar charging, and requires a Ring Home plan if you want to store your recordings. For more details of what you get with Ring Home, see our guide do I need a Ring subscription.

The price difference between the wireless Swann and Ring cameras disappears immediately if you want solar charging, since Ring's solar panels start at $39.99 / £39.99 / AU$59 each.

We're hoping to test the Swann EVO Wireless Solar soon, to see how it compares with other wireless cams in its price bracket and above.

You might also like

• How to install a smart lock in six simple steps
• Roku's new wireless security cameras will guard your home for up to two years on a single charge
• Eufy's new smart lock reads the veins in your palm for touchless entry – even if your hands are dirty

As artificial intelligence (AI) tools like ChatGPT, Co-Pilot, Grok and predictive analytics platforms become embedded in everyday business operations, many companies are unknowingly walking a legal tightrope.

While the potential of AI tools provide many benefits - streamlining workflows, enhancing decision-making, and unlocking new efficiencies - the legal implications are vast, complex, and often misunderstood.

From data scraping to automated decision-making, the deployment of AI systems raises serious questions around copyright, data protection, and regulatory compliance.

Without robust internal frameworks and a clear understanding of the legal landscape, businesses risk breaching key laws and exposing themselves to reputational and financial harm.

GDPR and the Use of AI on Employee Data

One of the most pressing concerns is how AI is being used internally, particularly when it comes to processing employee data. Many organizations are turning to AI to support HR functions, monitor productivity, or even assess performance. However, these applications may be in direct conflict with the UK General Data Protection Regulation (GDPR).

GDPR principles such as fairness, transparency, and purpose limitation are often overlooked in the rush to adopt new technologies. For example, if an AI system is used for employee monitoring without their informed consent, or if the data collected is repurposed beyond its original intent, the business could be in breach of data protection law.

Moreover, automated decision-making that significantly affects individuals, such as hiring or disciplinary actions, requires specific safeguards under GDPR, including the right to human intervention.

The Legal Grey Area of Data Scraping

Another legal minefield is the use of scraped data to train AI models. While publicly available data may seem fair game, the reality is far more nuanced. Many websites explicitly prohibit scraping in their terms of service, and using such data without permission can lead to claims of breach of contract or even copyright infringement.

This issue is particularly relevant for businesses developing or fine-tuning their own AI models. If training data includes copyrighted material or personal information obtained without consent, the resulting model could be tainted from a legal standpoint. Even if the data was scraped by a third-party vendor, the business using the model could still be held liable.

Copyright Risks in Generative AI

Generative AI tools, such as large language models and image generators, present another set of challenges. Employees may use these tools to draft reports, create marketing content, or process third-party materials. However, if the input or output involves copyrighted content, and there are no proper permissions or frameworks in place, the business could be at risk of infringement.

For instance, using generative AI to summarize or repurpose a copyrighted article without a license could violate copyright law. Similarly, sharing AI-generated content that closely resembles protected work may also raise legal red flags. Businesses must ensure their employees understand these limitations and are trained to use AI tools within the bounds of copyright law.

The Danger of AI “Hallucinations”

One of the lesser-known but increasingly problematic risks of AI is the phenomenon of “hallucinations”- where AI systems generate outputs that are factually incorrect or misleading, but presented with confidence. In a business context, this can have serious consequences.

Consider a scenario where an AI tool is used to draft a public document or legal summary, in which it includes fabricated company information or incorrect regulations. If that content is published or relied upon, the business could face reputational damage, client dissatisfaction, or even legal liability. The risk is compounded when employees assume the AI’s output is accurate without proper verification.

The Need for Internal AI Governance

To mitigate these risks, businesses must act promptly to implement robust internal governance frameworks. This includes clear policies on how AI tools can be used, mandatory training for employees, and regular audits of AI-generated content.

Data Protection Impact Assessments (DPIAs) should be conducted whenever AI is used to process personal data, and ethical design principles should be embedded into any AI development process.

It’s also critical to establish boundaries around the use of proprietary or sensitive information. Employees interacting with large language models must be made aware that anything they input could potentially be stored or used to train future models. Without proper safeguards, there’s a real risk of inadvertently disclosing trade secrets or confidential data.

Regulatory Focus in 2025

Regulators are increasingly turning their attention to AI. In the UK, the Information Commissioner’s Office (ICO) has made it clear that AI systems must comply with existing data protection laws, and it is actively investigating cases where this may not be happening. The ICO is particularly focused on transparency, accountability, and the rights of individuals affected by automated decision-making.

Looking ahead, we can expect more guidance and enforcement around the use of AI in business. The UK is currently consulting on its AI Bill which aims to regulate artificial intelligence by establishing an AI Authority, enforcing ethical standards, ensuring transparency, and promoting safe, fair, and accountable AI development and use that businesses must comply with.

AI is transforming the way we work, but it’s not a free pass to bypass legal and ethical standards. Businesses must approach AI adoption with caution, clarity, and compliance to safeguard their staff and reputation. By investing in governance, training, and legal oversight, organizations can harness the power of AI while avoiding the pitfalls.

The legal risks are real, but with the right approach, they are also manageable.

We feature the best cloud document storage.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro