[{"data":1,"prerenderedAt":489},["ShallowReactive",2],{"blog-list-en":3},[4,114,252,378,421],{"id":5,"title":6,"body":7,"cover":94,"coverAlt":95,"date":96,"description":97,"draft":98,"extension":99,"lang":100,"linkedinUrl":84,"meta":101,"navigation":102,"path":103,"seo":104,"slug":105,"stem":106,"tags":107,"translationKey":111,"updated":112,"__hash__":113},"blog\u002Fblog\u002Fen\u002Fdeskilling-misses-the-point.md","Deskilling Misses the Point",{"type":8,"value":9,"toc":88},"minimark",[10,14,17,22,25,28,31,34,37,43,47,58,61,64,73,76,79],[11,12,13],"p",{},"So here we have deskilling, as people call it these days — a bit of a nuisance. It is another observation from the October “AI Transformation of Healthcare” conference. This time, the fear that we will forget how to think if we just click through AI results.",[11,15,16],{},"I would probably write myself off socially if I claimed that the concern was unfounded. I am just convinced that it misses the point.",[18,19,21],"h2",{"id":20},"the-problem-is-the-volume-of-work-not-the-loss-of-skill","The problem is the volume of work, not the loss of skill",[11,23,24],{},"The problem is the volume of routine work we cannot get through at a decent quality. Checking, comparing, judging similar cases over and over. If I can no longer simplify those blasted processes and strip out the steps that can be stripped out, then why not bring in AI — which, this time, is not the thing being cursed?",[11,26,27],{},"And what is worse, what if the repetitive activity is a key part of the process — an imaging examination, for instance? Rows of images, similar tasks, attention that has to last the whole shift. Is the problem really that, with the help of AI, I start to lose my skills? Or is it rather that our brain does not deliver the machine-like stability of performance?",[11,29,30],{},"Human performance fluctuates. We get tired, we lose focus, towards the end of the day we see things differently than at the start. That is not a moral failure; it is a property of the brain. A machine, in this sense, does not get bored and does not wander. I am not saying it is infallible. I am saying that its error rate can be described, and the human one usually cannot. That is why it seems a shame to me to fear the tool before we know how a person performs at the same work.",[11,32,33],{},"Enough questions. The whole thing comes down to this: with AI, we can measure the reliability of the output. And when we do it properly, we know, transparently and repeatably, how it performs. Fine, we can debate the quality of the dataset, what the model was tested on and how far that carries over into practice. About a person, we do not know. We ask about education, experience and the impression that they are “good”. That is not the same kind of information as reliability measured over and over. We only guess, believe, hope or fear — but we usually do not quantify the quality of human output.",[11,35,36],{},"When we then put AI, for which we have a number, next to a person, for whom we do not, the person can look worse even if they have not changed. So if it seems somewhere that the quality of human output falls when AI is used, the problem will not be, in principle, with the AI. It is more that there was simply no measure of the quality of human output. AI merely makes visible what we have been living with, not how we have got worse.",[38,39],"blog-figure",{"alt":40,"caption":41,"src":42},"A two-panel satirical illustration: on the left, a content official in pink-tinted glasses stamps papers at a messy desk; on the right, three people in business attire point at a crying AI terminal with a gauge.","With a machine, we measure the error rate and point the finger at it. With a person, we usually do not measure it at all.","\u002Fimages\u002Fblog\u002Fdeskilling\u002Fen\u002Fchybovost.webp",[18,44,46],{"id":45},"when-human-work-really-is-measured","When human work really is measured",[11,48,49,50,57],{},"It is no accident that the debate so often turns on imaging. Repetitive reading is the core of the work there — and at the same time one of the few places where human performance can be compared with a number at all. A typical example is colonoscopy. The indicator is the adenoma detection rate: the share of examinations in which the doctor finds at least one adenoma. Precisely because that number exists, it was possible to find something out at all when, in August 2025, an observational study from Poland was published in ",[51,52,56],"a",{"href":53,"rel":54},"https:\u002F\u002Fwww.thelancet.com\u002Fjournals\u002Flangas\u002Farticle\u002FPIIS2468-1253(25)00133-5\u002Fabstract",[55],"nofollow","Lancet Gastroenterology & Hepatology",".",[11,59,60],{},"Nineteen doctors at four centres were compared over three months before a polyp-detection system was introduced and three months after it — always in examinations without AI. The question, then, was not how a person performs with AI, but how the same doctor performs without it after growing used to it in the rest of their work. The adenoma detection rate fell from 28.4% to 22.4%. That is the result now cited as evidence of deskilling.",[11,62,63],{},"I would be cautious about it, though. It is an observational comparison, not a randomized study. The patient groups differed, and examination time, which strongly affects detection, was not tracked. The number of adenomas per examination, and the detection of advanced findings and of cancer, did not change statistically. The authors themselves treat it as a hypothesis to be tested, and call for a proper prospective study.",[11,65,66,67,72],{},"And above all: what matters for the patient is how well the person and the AI do it together. There, randomized studies repeatedly show that detection rises. A meta-analysis of ",[51,68,71],{"href":69,"rel":70},"https:\u002F\u002Fwww.giejournal.org\u002Farticle\u002FS0016-5107(24)03471-0\u002Ffulltext",[55],"28 randomized studies with nearly 24,000 patients",", published in 2024 in Gastrointestinal Endoscopy, reports that with AI the adenoma detection rate was 20% higher — mainly for smaller lesions; for advanced adenomas, no difference appeared. If we are afraid that skill will dull without AI, that is an argument for measuring people regularly, not an argument against a tool that improves detection when they work together.",[11,74,75],{},"A measure does not have to be confined to healthcare. Where the output can be compared with the facts, or with a second reading, human reliability can be tracked in the same way as a machine’s: on a sample, repeatedly, not once in an audit. Without that, deskilling is easy to talk about and hard to judge. Not so that we stop trusting people, but so that we do not have to argue it out from our desks. First the number, then the conclusion: whether we really are getting worse, or whether we are only now seeing what has been there all along.",[11,77,78],{},"Do you measure at all, where you are, how reliable human output is? Or do we only guess and hope?",[11,80,81,82,87],{},"I published a shorter version on ",[51,83,86],{"href":84,"rel":85},"https:\u002F\u002Fwww.linkedin.com\u002Ffeed\u002Fupdate\u002Furn:li:activity:7523303803706613761\u002F",[55],"LinkedIn"," on 3 November 2026.",{"title":89,"searchDepth":90,"depth":90,"links":91},"",2,[92,93],{"id":20,"depth":90,"text":21},{"id":45,"depth":90,"text":46},"\u002Fimages\u002Fblog\u002Fdeskilling\u002Fcover.webp","A schematic orange line drawing on a dark background: on the left, a figure at a laptop breaks into dashed lines with a fluctuating curve running through it; on the right, an empty gauge circle with no needle.","2026-11-03","The fear that we will forget how to think if we just click through AI results is fair. But it misses the point: the real problem is that we usually do not measure the quality of human output at all.",false,"md","en",{},true,"\u002Fblog\u002Fen\u002Fdeskilling-misses-the-point",{"title":6,"description":97},"deskilling-misses-the-point","blog\u002Fen\u002Fdeskilling-misses-the-point",[108,109,110],"AI","Deskilling","Digitalization","deskilling",null,"oAhAWwLJfdVTDdsTE_bo54tuSJwNZ_nASM6L_y_XKyQ",{"id":115,"title":116,"body":117,"cover":239,"coverAlt":240,"date":241,"description":242,"draft":98,"extension":99,"lang":100,"linkedinUrl":232,"meta":243,"navigation":102,"path":244,"seo":245,"slug":246,"stem":247,"tags":248,"translationKey":250,"updated":112,"__hash__":251},"blog\u002Fblog\u002Fen\u002Fwhat-to-fear-with-ai.md","What to Fear with AI – and What Not To",{"type":8,"value":118,"toc":235},[119,122,125,128,136,151,156,159,168,172,175,195,198,203,207,210,213,216,219,222,225,228],[11,120,121],{},"The subject of data protection and AI leaves me a bit grouchy. It is another observation from the October “AI Transformation of Healthcare” conference. We fear something that almost never happens in practice, we overlook what actually can happen, and we fail to see what we are missing out on.",[11,123,124],{},"The risk that gets repeated over and over is that sensitive data will become part of a model’s knowledge and someone will extract them from it. That concern is debatable. A language model can be seen as a form of lossy compression: pulling specific training data back out, one-to-one, is rare and takes a specialized attack.",[11,126,127],{},"In our “cloud laboratories” — read: on a single virtual machine — we tried this on 25 October 2026 with Llama 3 8B. No chat template, temperature set to zero, so the result could be reproduced.",[11,129,130,131,135],{},"First we gave the model the beginning of a sentence from the U.S. Declaration of Independence, a text that exists in countless copies on the internet. It completed it word for word, all 48 words. Then the opening of Mácha’s ",[132,133,134],"em",{},"Máj",", a famous Czech romantic poem by K. H. Mácha, well known here but far rarer in global training data. It did not finish the poem. It went its own way: after “první máj” (“the first of May”) came “první červen – první červenec” (“the first of June – the first of July”). Finally, the start of a fictional discharge report the model could never have seen. It invented a diagnosis and then got stuck in a loop on a date.",[11,137,138,139,144,145,150],{},"This is not an accident. Meta reports ",[51,140,143],{"href":141,"rel":142},"https:\u002F\u002Fai.meta.com\u002Fblog\u002Fmeta-llama-3\u002F",[55],"over 15 trillion training tokens"," for Llama 3; the weights of the 8-billion-parameter model come to around 16 GB, three orders of magnitude less than the text itself would occupy. Most of the verbatim information therefore has to be discarded. This is not proof that a model remembers nothing. It is a demonstration. It recalls verbatim mainly what it has seen many times. A unique record it will not repeat; it invents its own — and that is a different risk, because the output can look like a genuine document. Targeted attacks in the literature have managed to ",[51,146,149],{"href":147,"rel":148},"https:\u002F\u002Fwww.usenix.org\u002Fconference\u002Fusenixsecurity21\u002Fpresentation\u002Fcarlini-extracting",[55],"extract even rarer texts"," from models, sometimes including personal data.",[38,152],{"alt":153,"caption":154,"src":155},"Infographic from a Llama 3 8B test: the Declaration of Independence completed verbatim, Mácha’s Máj not completed, and an invented diagnosis for a fictional discharge report.","Our test of 25 October 2026 on Llama 3 8B. Only a frequently repeated text was recalled verbatim; for the fictional report, the model invented a diagnosis.","\u002Fimages\u002Fblog\u002Fdata-a-bezpecnost\u002Fen\u002Fwhat-ai-remembers.webp",[11,157,158],{},"The greater risk is therefore the flow of data itself, and its short-term storage at the provider, should a breach and a leak occur there.",[11,160,161,162,167],{},"But the enterprise services of the major providers do not, as a rule, train on customer data, and retention can be limited by contract, with a number of services down to zero. That is precisely so they can minimize risk and ease customers’ concerns. Running models locally may look like the answer. But which of us has our own security team, patches around the clock and watches access better than Microsoft or Google? According to the ",[51,163,166],{"href":164,"rel":165},"https:\u002F\u002Fwww.ibm.com\u002Freports\u002Fdata-breach",[55],"IBM Cost of a Data Breach 2026"," report, 30% of breaches involved data in companies’ own infrastructure, and 23% in the public cloud. On-premises infrastructure is not automatically safer.",[18,169,171],{"id":170},"what-we-lose-when-we-do-not-use-ai","What we lose when we do not use AI",[11,173,174],{},"I would like the debate to turn towards how we map and use the options we probably already have in the company — typically Microsoft Azure, Google Vertex or AWS — or how we find and contract AI services at enterprise level. It would help to start with the question of what we lose when we do not use AI, and not only with everything that might go wrong.",[11,176,177,178,183,184,189,190,194],{},"Every day that we do not have AI, or cannot have it, someone else in our field is profiting from it. According to ",[51,179,182],{"href":180,"rel":181},"https:\u002F\u002Fec.europa.eu\u002Feurostat\u002Fstatistics-explained\u002Findex.php?title=Use_of_artificial_intelligence_in_enterprises",[55],"Eurostat",", in 2025 at least one AI technology was used by 17.6% of enterprises with 10 or more employees in the Czech Republic and by 42% in Denmark (dataset ",[51,185,188],{"href":186,"rel":187},"https:\u002F\u002Fec.europa.eu\u002Feurostat\u002Fdatabrowser\u002Fproduct\u002Fpage\u002Fisoc_eb_ai",[55],"isoc_eb_ai","). The same GDPR applies in both. Among EU firms that had considered AI but do not use it, ",[51,191,193],{"href":186,"rel":192},[55],"52.7%"," cite concerns about data protection and privacy. At best, then, we are not developing our own potential; at worst, over time, it could cost us our existence.",[11,196,197],{},"And once something as useful as AI is available, people will find a way to benefit from it anyway. Then you have shadow AI, ready-made.",[38,199],{"alt":200,"caption":201,"src":202},"Infographic ‘What to fear with AI – and what not to’: the fear of data being extracted from a model, the real risk of data flowing to the provider, and the loss from not using AI.","Three layers of the debate: what we fear, what can actually happen, and what we miss out on when we do not use AI.","\u002Fimages\u002Fblog\u002Fdata-a-bezpecnost\u002Fen\u002Fwhat-to-fear.webp",[18,204,206],{"id":205},"which-data-belongs-where","Which data belongs where",[11,208,209],{},"It does not have to be complicated. Just three levels, and a clear rule for what may go to each.",[11,211,212],{},"A public service, typically a private account in a browser, is no place for patients’ or customers’ personal data, trade secrets, credentials or sensitive internal documents. A general description of a problem, or a public text, often is. This is exactly where people will turn if the company offers them nothing else.",[11,214,215],{},"An enterprise cloud service, ideally with storage in the EU, contractually barred from training on customer data and with short or zero retention, can already take ordinary internal work: drafts, analyses, summaries. Not everything belongs there either. But it is a place where the risk of data flowing out can be handled by contract and configuration, not by a ban.",[11,217,218],{},"A local model on your own machine or on your own network makes sense where the data should not leave the company at all. Not because it is automatically safer, but because the data do not travel outside. Patching, access and operations are then our responsibility.",[11,220,221],{},"That also includes an agreement on who will check the result before it is used for real. Responsibility stays with the person, not the model.",[11,223,224],{},"So the question is not whether to be afraid of data, but how to sort them when working with AI.",[11,226,227],{},"The obligatory question at the end remains the same: where you are, are you still asking “whether to use AI at all”, or already “which kind, and on what terms”?",[11,229,81,230,234],{},[51,231,86],{"href":232,"rel":233},"https:\u002F\u002Fwww.linkedin.com\u002Ffeed\u002Fupdate\u002Furn:li:activity:7521543564921659392\u002F",[55]," on 29 October 2026.",{"title":89,"searchDepth":90,"depth":90,"links":236},[237,238],{"id":170,"depth":90,"text":171},{"id":205,"depth":90,"text":206},"\u002Fimages\u002Fblog\u002Fdata-a-bezpecnost\u002Fcover.webp","A schematic view of data protection when using AI: data flows, a local model and an enterprise service.","2026-10-29","We fear that sensitive data will become part of the model. The greater risk is the flow of data to the provider — and a still greater loss if we do not use AI.",{},"\u002Fblog\u002Fen\u002Fwhat-to-fear-with-ai",{"title":116,"description":242},"what-to-fear-with-ai","blog\u002Fen\u002Fwhat-to-fear-with-ai",[108,249,110],"Data Protection","data-a-bezpecnost","bZwWDVUiXrxOUROHTdUwOv_XKGD6wJOrxrXIJRF5kik",{"id":253,"title":254,"body":255,"cover":365,"coverAlt":366,"date":367,"description":368,"draft":98,"extension":99,"lang":100,"linkedinUrl":357,"meta":369,"navigation":102,"path":370,"seo":371,"slug":372,"stem":373,"tags":374,"translationKey":376,"updated":112,"__hash__":377},"blog\u002Fblog\u002Fen\u002Ftwo-layers-of-ai.md","Two Layers of AI",{"type":8,"value":256,"toc":360},[257,260,263,266,270,273,276,279,282,297,300,304,307,322,325,328,332,335,338,341,344,347,350,353],[11,258,259],{},"In October 2026 I attended the “AI Transformation of Healthcare” conference, organized by the Czech Association of Artificial Intelligence (ČAUI). One idea took shape there so clearly that I wrote it up on LinkedIn. I want to add to it here.",[11,261,262],{},"Today, almost all talk about artificial intelligence is about products: how to choose them, roll them out, pay for them and run them safely. That layer is important. And it is mostly about the difficulties — not in the product itself, but around it. Typically integration into existing systems, responsibility for decisions, data protection, who will pay for it once the pilot ends, and above all how people’s work changes and whether they trust it.",[11,264,265],{},"Top-down rollouts are slow and expensive not because of the technology, but because of the organization and its culture. That is as true in a hospital as it is in a factory.",[18,267,269],{"id":268},"the-second-layer-is-harder-to-see","The second layer is harder to see",[11,271,272],{},"AI puts into the hands of individuals something that used to belong only to teams of developers, analysts and data scientists: the chance to turn their own curiosity into a working solution. The person who knows the problem best can now build an application, a predictive model or a data analysis themselves, in days rather than months. In principle, their idea no longer has to wait for a budget, a tender or spare capacity in IT.",[11,274,275],{},"I don’t mean large programs. I mean things that start at someone’s desk. A quality engineer puts together a picture of recurring defects from photos of faults. A process engineer uses weekly outputs to spot when a machine starts to behave differently. A doctor drafts a checklist from professional guidelines and shares it with colleagues for comment. None of this is an “AI project”. It is someone who sees the problem every day and suddenly has a tool to try something about it.",[11,277,278],{},"That, to me, is the real change: who gets to innovate. Innovation stops being the preserve of specialist teams and appears where the problem actually is, so it sidesteps the barriers that stall a rollout from the top.",[11,280,281],{},"People are already doing this. Often quietly.",[11,283,284,285,290,291,296],{},"According to the ",[51,286,289],{"href":287,"rel":288},"https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fworklab\u002Fwork-trend-index\u002Fai-at-work-is-here-now-comes-the-hard-part",[55],"2024 Work Trend Index survey",", commissioned by Microsoft and LinkedIn among 31,000 knowledge workers in 31 countries, 78 % of AI users bring their own tools to work. The survey calls this BYOAI, bring your own AI. It happens without guidance or approval from the top. ",[51,292,295],{"href":293,"rel":294},"https:\u002F\u002Fassets-c4akfrf5b4d3f4b7.z01.azurefd.net\u002Fassets\u002F2024\u002F05\u002F2024_Work_Trend_Index_Annual_Report_6_7_24_666b2e2fafceb.pdf",[55],"In the Czech Republic it was 86 %"," of AI users, and 80 % across Europe as a whole. The same survey finds that 52 % of people who use AI at work are reluctant to admit they use it for their most important tasks.",[11,298,299],{},"That is not an ideal state of affairs: if a company does not create the space, people will find it anyway — just without rules, without sharing, and often with data that should never have ended up in a public tool.",[18,301,303],{"id":302},"a-story-that-is-repeating-itself","A story that is repeating itself",[11,305,306],{},"Perhaps innovation from the ground up is simply unstoppable. One legendary anecdote will do.",[11,308,309,310,315,316,321],{},"In the late 1970s the first spreadsheet application, VisiCalc, by Dan Bricklin and Bob Frankston, came out for the Apple II. The ",[51,311,314],{"href":312,"rel":313},"https:\u002F\u002Fwww.computerhistory.org\u002Ftimeline\u002Fsoftware-languages\u002F",[55],"Computer History Museum"," calls it the program that turned the personal computer into a business machine. Shops demonstrated it on the Apple II, and people from companies, often in finance and accounting, added the computer and the software to the order and started calculating at their own desks; ",[51,317,320],{"href":318,"rel":319},"https:\u002F\u002Fwww.theguardian.com\u002Fmedia\u002F2009\u002Fjan\u002F18\u002Fvisicalc-apple-software-bricklin",[55],"as John Naughton described it",", it sometimes went on expenses. They could now try for themselves the analyses they used to send to the data-processing department. A few years later, there was a spreadsheet in every office.",[11,323,324],{},"I think the same thing is happening with AI, only much faster.",[11,326,327],{},"At the same time, I see no reason why the two layers should be mutually exclusive. Products bring scale, reliability and clearly assigned responsibility. Building from the ground up brings speed and knowledge of a specific problem. So the question is not whether to deploy AI, but how to create an environment where people can build, and take their ideas safely all the way to real use.",[18,329,331],{"id":330},"how-to-create-an-environment-for-building-from-the-ground-up","How to create an environment for building from the ground up",[11,333,334],{},"It does not have to be an elaborate program. Four plain rules will do.",[11,336,337],{},"First, a protected space to try things out. People need approved tools in which they can experiment. Otherwise they end up on a private account in a public service — precisely the situation companies fear, and often bring on themselves with a ban.",[11,339,340],{},"Second, clear rules for data. Patient or customer personal data, trade secrets, credentials and sensitive internal documents do not belong outside; a general description of a problem or an anonymized sample often does. People will follow the rule if they know it and it is short. The worst thing is silence, because in silence everyone sets their own threshold.",[11,342,343],{},"Third, guardrails. An experiment is not production: the result of one person’s workshop should not by itself run a line, approve a batch release, or go into a medical record. Before it is used for real, it must be clear who will check it — responsibility stays with the person, not the model.",[11,345,346],{},"Fourth, sharing. When something genuinely helps someone, there should be a short way to show it to others and, if appropriate, turn it into a supported tool. Otherwise good ideas stay at one desk, and risky ones still spread quietly.",[11,348,349],{},"None of this will stop innovation from the ground up. It can, however, separate it from unnecessary risk and give it a chance to grow.",[11,351,352],{},"The obligatory question at the end: how do you see it where you are? Does it make more sense to scale finished products — “off-the-shelf boxes”, if you like — or to give people room to build from the ground up? I lean towards a combination of the two.",[11,354,81,355,359],{},[51,356,86],{"href":357,"rel":358},"https:\u002F\u002Fwww.linkedin.com\u002Ffeed\u002Fupdate\u002Furn:li:activity:7517893845146226688\u002F",[55]," on 19 October 2026.",{"title":89,"searchDepth":90,"depth":90,"links":361},[362,363,364],{"id":268,"depth":90,"text":269},{"id":302,"depth":90,"text":303},{"id":330,"depth":90,"text":331},"\u002Fimages\u002Fblog\u002Fdve-roviny-ai\u002Fcover.webp","A schematic view of two layers of AI: rolling out products from the top, and building from the ground up.","2026-10-19","AI is mostly treated as a product to roll out. The second layer changes who can innovate. How to let people build safely from the ground up.",{},"\u002Fblog\u002Fen\u002Ftwo-layers-of-ai",{"title":254,"description":368},"two-layers-of-ai","blog\u002Fen\u002Ftwo-layers-of-ai",[108,375,110],"Innovation","dve-roviny-ai","AZTZD6mECubaGOsmKXXnCwVuV1VU5FrWETySh5jZuhc",{"id":379,"title":380,"body":381,"cover":390,"coverAlt":388,"date":409,"description":410,"draft":98,"extension":99,"lang":100,"linkedinUrl":112,"meta":411,"navigation":102,"path":412,"seo":413,"slug":414,"stem":415,"tags":416,"translationKey":419,"updated":112,"__hash__":420},"blog\u002Fblog\u002Fen\u002Fsolidrun-neuralcam-partnership.md","SolidRun: Nitritty Labs Empowers Immobile Patients with NeuralCam",{"type":8,"value":382,"toc":407},[383,386,391,394,400],[11,384,385],{},"Check out the latest article from our partner SolidRun, which presents our AI camera NeuralCam, powered by SolidRun hardware with the Hailo-15 AI processor. NeuralCam enables immobile patients to call for help with a simple eye blink.",[38,387],{"alt":388,"caption":389,"src":390},"SolidRun partner logo","NeuralCam runs on SolidRun hardware with the Hailo-15 AI processor.","\u002Fimages\u002Fblog\u002Fsolidrun-neuralcam-partnership\u002Fcover.png",[11,392,393],{},"Nitritty Labs empowers immobile patients with NeuralCam — powered by SolidRun’s Hailo-15 SoM.",[11,395,396],{},[51,397,399],{"href":398},"\u002Fdocuments\u002FSolidRun-NeuralCam-Hailo-15.pdf","Download SolidRun feature PDF",[11,401,402,403,57],{},"Learn more about ",[51,404,406],{"href":405},"\u002Fneuralcam","NeuralCam",{"title":89,"searchDepth":90,"depth":90,"links":408},[],"2025-12-31","Check out the latest article from our partner SolidRun, which presents our AI camera NeuralCam, powered by SolidRun hardware with the Hailo-15 AI processor.",{},"\u002Fblog\u002Fen\u002Fsolidrun-neuralcam-partnership",{"title":380,"description":410},"solidrun-neuralcam-partnership","blog\u002Fen\u002Fsolidrun-neuralcam-partnership",[406,417,418],"SolidRun","Partnership","solidrun","ONtSvFoYZd5Ionsh1FKBJNeS4HW-m86r7i1yeXN3RMw",{"id":422,"title":423,"body":424,"cover":433,"coverAlt":476,"date":477,"description":478,"draft":98,"extension":99,"lang":100,"linkedinUrl":112,"meta":479,"navigation":102,"path":480,"seo":481,"slug":482,"stem":483,"tags":484,"translationKey":487,"updated":112,"__hash__":488},"blog\u002Fblog\u002Fen\u002Fneuralcam-assistive-camera.md","Nitritty NeuralCam: AI Camera for Patients Gives Immobile Individuals a Way to Call for Help",{"type":8,"value":425,"toc":469},[426,429,434,438,441,445,448,452,455,459,462,466],[11,427,428],{},"Nitritty Labs has developed NeuralCam, an AI-powered assistive camera for patients that enables fully or partially immobile patients to request assistance using controlled eye blinks. The solution addresses a critical challenge in healthcare environments, where patients with paralysis, limited mobility, or speech impairments may be unable to use conventional nurse-call systems.",[38,430],{"alt":431,"caption":432,"src":433},"NeuralCam assistive camera in a hospital room","NeuralCam uses on-device eye tracking so patients can request help with a blink.","\u002Fimages\u002Fblog\u002Fneuralcam-assistive-camera\u002Fcover.png",[18,435,437],{"id":436},"reliable-assistance-when-it-matters-most","Reliable Assistance When It Matters Most",[11,439,440],{},"NeuralCam continuously monitors the patient’s eye state and recognizes intentional blinking patterns that indicate a request for help. Once confirmed, the system immediately triggers a signal to the existing nurse-call infrastructure, allowing caregivers to respond promptly and reliably. Patients who previously depended on frequent manual checks can now summon assistance independently, reducing response delays and improving overall safety.",[18,442,444],{"id":443},"designed-for-accuracy-privacy-and-trust","Designed for Accuracy, Privacy, and Trust",[11,446,447],{},"The device delivers consistent eye recognition across all lighting conditions, including low-light and nighttime environments. All video processing is performed directly inside the camera, with no transmission, storage, or external sharing of image data. This on-device approach ensures full patient privacy, minimizes risk, and complies with GDPR and strict medical data protection requirements.",[18,449,451],{"id":450},"edge-ai-at-the-core","Edge AI at the Core",[11,453,454],{},"NeuralCam is built on an edge AI platform combining a high-performance AI vision processor with a modern image sensor. This hardware enables real-time eye-state analysis with ultra-low latency, high accuracy, and energy-efficient operation.",[18,456,458],{"id":457},"seamless-integration-into-care-environments","Seamless Integration into Care Environments",[11,460,461],{},"The solution integrates smoothly with existing patient call systems and requires minimal training for medical staff. A clear visual feedback indicator confirms successful recognition of a help request, giving patients confidence that their signal has been received. Fast installation and commissioning make the system suitable for hospitals, rehabilitation centers, and assisted living facilities.",[18,463,465],{"id":464},"a-small-gesture-with-a-big-impact","A Small Gesture with a Big Impact",[11,467,468],{},"NeuralCam restores a vital communication channel to patients who would otherwise struggle to call for help. By combining edge AI, privacy-first design, and seamless integration into clinical workflows, the solution enhances patient safety, supports caregivers, and preserves dignity—demonstrating how even a simple blink can make a meaningful difference in healthcare.",{"title":89,"searchDepth":90,"depth":90,"links":470},[471,472,473,474,475],{"id":436,"depth":90,"text":437},{"id":443,"depth":90,"text":444},{"id":450,"depth":90,"text":451},{"id":457,"depth":90,"text":458},{"id":464,"depth":90,"text":465},"NeuralCam assistive camera for patient care","2025-12-29","Nitritty Labs has developed NeuralCam, an AI-powered assistive camera for patients that enables fully or partially immobile patients to request assistance using controlled eye blinks.",{},"\u002Fblog\u002Fen\u002Fneuralcam-assistive-camera",{"title":423,"description":478},"neuralcam-assistive-camera","blog\u002Fen\u002Fneuralcam-assistive-camera",[406,485,486],"Healthcare","Edge AI","neuralcam","29aj2kPAb5C3velfYJxvFn_c0A9R2fP9edFV8GSIZOw",1794457154012]