We unlock our phones with our faces, dictate messages to smart speakers, and get alerts from our smart locks when the delivery driver rings the bell. These seemingly ordinary actions rely on layers of technology that are evolving rapidly, while the rules of the game are changing just as quickly. Several tech trends are reshaping our everyday digital habits, with tangible consequences for security, energy consumption, and the way companies handle our data.
The AI Act and algorithmic transparency: what’s changing for users in 2026
When we use a customer service chatbot or an image-generation tool, we interact with artificial intelligence without always knowing it. The AI Act, the European Union’s regulation on AI, is beginning to change that. Since February 2025, certain practices have been banned (social scoring, subliminal manipulation), and an obligation to develop AI literacy has come into force.
The timeline is becoming clearer: from August 2, 2026, AI systems will have to clearly disclose that users are interacting with a machine. Synthetic content, including deepfakes, will be subject to disclosure rules. Requirements for general-purpose AI models have applied since August 2025.
In practice, we’re keeping an increasingly close eye on developments in these regulations through Ei Mag’s tech section, because the gap between what the law requires and what platforms actually implement remains something to watch month after month.
For companies integrating AI into their services (product recommendations, CV screening, customer scoring), classification as “high risk” brings significant documentation requirements. The timeline for high-risk uses has also recently been pushed back, creating a period of regulatory uncertainty for the technical teams responsible for compliance.
Digital carbon footprint: devices have a greater impact than the cloud
We often assume that digital pollution comes from data centers or video streaming. The updated ADEME study published in January 2025 shifts the focus: devices are the largest contributor to France’s digital carbon footprint, mainly because of their manufacture. Smartphones, laptops, tablets, connected devices: extracting raw materials and assembling products account for the bulk of the impact, long before they are used.
Data centers also account for a significant share of this footprint. The proliferation of AI services, which require considerable computing power, is putting additional pressure on this infrastructure.
Reducing your impact
Extending a device’s lifespan remains the most direct way to reduce its impact. Keeping a smartphone for one extra year cuts its footprint more than choosing a “green” plan or offsetting emissions with carbon credits.
- Choose refurbished computers and smartphones, checking the warranty and battery capacity.
- Limit the replacement of connected devices (watches, speakers, cameras) to cases where the hardware no longer receives security updates.
- Consolidate cloud services rather than signing up for multiple platforms that duplicate data storage.
Everyday cybersecurity: the threat has shifted targets
Cybersecurity technologies are advancing, but so are attacks. We’re seeing a shift: phishing campaigns no longer target only work email accounts. They also target individuals with text messages impersonating delivery services, banks, or government agencies. Personal smartphones have become the weak link in digital security.
The Internet of Things makes the situation even more complicated. A poorly configured connected camera or a thermostat whose firmware is never updated: every device added to the home network creates a potential entry point.
Practical steps to secure your digital habits
- Enable two-factor authentication on every account that offers it, preferably using an authentication app rather than a code sent by text.
- Check the permissions granted to apps: a smart light bulb doesn’t need access to your microphone.
- Segment your home Wi-Fi network by creating a dedicated network for connected devices, separate from the one used for work or banking.
On-device AI: processing without the cloud
AI doesn’t run only on remote servers. More and more processing takes place directly on the device, an approach known as edge AI or on-device AI. A phone’s facial recognition, offline voice transcription, fall detection on a smartwatch: these features run locally without sending data to a server.
This approach reduces latency and the volume of data transferred, and to some extent, the risks to privacy. Sensitive data (biometrics, conversations) stays on the device instead of passing through a data center.
For businesses, on-device AI is changing how products are designed. An industrial sensor that can analyze vibrations in real time can detect an anomaly without waiting for a response from a remote server. Local processing speeds up decision-making and reduces reliance on the network.
The limitations depend on the complexity of the model and the power of the onboard processor: simple tasks (image classification, basic voice commands) work well locally, while large generative models still require server resources.
Today’s technology trends aren’t limited to spectacular announcements. They show up in regulatory requirements such as the AI Act, the environmental impact of our devices, security vulnerabilities in our connected devices, and the shift of computing to our own hardware. Keeping up with these developments means understanding what’s really changing in our everyday digital habits.
