Up Times • Feb. 2025

Among DeepSeek’s issues, it can be jailbroken easily (compared to other LLMs [e.g. ChatGPT]) – it lacks guardrails that might hinder someone from using it to generate new malware, cyberweapons (like Advanced Persistent Threats) and more.

Boston’s Enkrypt AI released a report about the Chinese DeepSeek-R1 AI chatbot and, maybe not too surprisingly, found the software lacking in security.

Our findings reveal that DeepSeek-R1’s security vulnerabilities could be turned into a dangerous tool – one that cyber criminals, disinformation networks and even those with biochemical warfare ambitions could exploit, [Sahil Agarwal Enkrypt AI’s CEO] said. These risks demand immediate attention … [6 min. read] computerweekly.com

Throughout my own university education, I didn’t have a single lecture on it … given that cybercrime is the single most common form of crime, this omission is unacceptable, Dr Julia Shaw

Dehuman

Psychological scientist Dr Julia Shaw begins to illuminate the differences between real-life crime and cyber crime: [To deter crime, you need to have] capable guardians … in cyberspace, authority is minimal and there is a perception that nobody is in charge. Because nobody is.

Cybercrime is [also] made easier because we can more readily dehumanise people online, Shaw explains, And when we stop seeing people as human beings, we may feel free to do more terrible things to them. To be online is to experience a disembodiment of ideas. The internet frees us from our physical selves, for better or for worse. And this leads to a flat experience, leaving behind the normal multi-sensory interaction we have with people in real life that reminds us that they are fleshy, vulnerable and sensitive … [9 min. read] wired.com

   

And the sheep will scatter

Bryley partner Huntress has released its 2025 Cyber Threat Report.

As far as Bryley clients are concerned, probably the most salient change observed by the threat-hunting organization comes in the wake of the FBI’s 2023 disruption of Qakbot ransomware. The group behind the main ransomware delivery mechanism was dismantled through US law enforcement efforts. But the net result has not been a long-term drop in ransomware attacks, but a number of newly created variants designed to evade detection.

Because of this Bryley advocates an Endpoint Detection and Response (EDR) security layer that complements traditional anti-malware by monitoring for suspicious activity on desktops and laptops, enabling automatic shutdown and removal of attackers … [1 hr. read] huntress.com

Dr Anna Lembke:I think people are unhappier than they’d like to be, and can’t figure out why.

Using

Psychiatrist Anna Lembke is an addiction specialist. For years she treated drug and alcohol addiction and then witnessed similar symptoms in the patterns people displayed in their tech use.

Essentially what’s happened is we’re spending more and more of our energy and creativity investing in this online world, which means that we are actually leaching our real-life existence of our energy and creativity. So when we try to re-enter the real world, it actually is more boring, because there’s less going on, because there’s nobody there … [18 min. read] nytimes.com [paywall]

Andrea Davis tells parents to teach kids: you have a choice … [you] don’t have to adopt every new technology that comes along

Andrea Davis of Better Screen Time is an advocate for raising kids to have a healthy relationship to technology as much as possible. In this AI-focused Facebook Reel, Davis advises that parents take responsibility to guide children to interact wisely with AI.

Though AI is presented in a friendly way, Davis says, its content is other people’s work and you need to do your own work. Also that AI makes mistakes and tells untruths, so you need to be critical of how it answers.

Davis says to encourage children to be suspicious of images online because of the presence of deepfakes. And re images, kids need to be taught the ethics of image sharing and the harm they can cause others.

Last, Davis makes clear that AI is a tool, not a friend. Though it is good at mimicking human relationships, it is no substitute. (I originally learned about Davis in the Wall Street Journal) [2 min. watch] facebook.com

^{Note: The section directly above is Bryley’s curated list of external stories. Bryley does not take credit for the content of these stories, nor does it endorse or imply an affiliation with the authors or publications in which they appear.}