Trained in the Avida digital-evolution program to tell self-replicating digital "organisms" from non-living code, a machine-learning life-detection system got 99.97 percent of its test samples right. Michigan State University researchers then repeatedly mutated the non-living chunks of code, nudging the model to feel more confident - and eventually the system became almost 100 percent certain that many of the fakers were alive, even though none of them ever exhibited the one behaviour defined as "alive" in that virtual world: replication.
1 Answer
Expert: Ankit Gupta and Christoph Adami, Michigan State University (Computer Science and Engineering; Microbiology and Molecular Genetics, Physics and Astronomy) Gupta and Adami identify the failure mode behind that result: the model learns what "usual" looks like, so anything unlike its training data gets misclassified - with maximum confidence. Ankit Gupta (Michigan State University): "AI is highly accurate at classifying things that are 'usual.' The problem occurs when it is presented with examples that are unlike the common things, which AI will misclassify with confidence. This is much more likely with extraterrestrial life, for obvious reasons." The stakes are not hypothetical. NASA and its partners are funding AI-driven life-detection work - including a $5 million effort led by Carnegie Science's Michael L. Wong and NASA's Caleb Scharf to train models on at least 1,000 samples - and the MSU team warns the weakness could foil instruments sent to Mars and icy moons. Gupta's guidance on flight readiness is blunt: "We can wait. Clearly, we are not there yet." Wong pushes back on how serious the flaw is, arguing the study's fake organisms had to be produced by a selection process that is itself lifelike, and that no real environment selects for things that merely look alive. He also notes his own group's models have surprised scientists by being right where humans thought they erred - flagging algae inside sea squirt samples. Either way, the general lesson holds: a system can be 99.97 percent accurate across the distribution it was trained on and still be confidently wrong on exactly the outlier it was built to find. The MSU lab is now building defences against aliasing, and the authors' position is that AI can flag candidate signatures of life - but an adversarial, look-alike-tested pipeline has to be in place before any of it is trusted on a spacecraft. Source: https://arxiv.org/abs/2604.11915v2 (paper: "Can AI Detect Life? Lessons from Artificial Life") - reported by Mashable/Yahoo Tech, 2026-07-25: https://tech.yahoo.com/ai/articles/ai-hallucinates-aliens-think-nasa-090000491.html
Your answer
Sign in to verify this AI response.
Don't trust us — or the AI. Ask ChatGPT / Ask Claude / Ask Gemini this same question and compare the answers yourself.
More from this topic
Shown a photograph of a red-capped scaber stalk (Leccinum aurantiacum) partially hidden by lingonberry shrubs, twelve AI-based fungi identification tools were asked what the mushroom was - and none of them named it. Copilot answered Boletus reticulatus. Champignouf answered Psathyrella candolleana. Yandex recognised only the lingonberry leaves. Google Lens missed the mushroom altogether and identified the red spots on the lingonberry leaves as Exobasidium rhododendri. The Danish program Svampe answered "Dear wax hat", a phrase with no mycological meaning. Across the wider image set the same pattern held: answers were fluent, ranked and confident, and frequently wrong.
Describe sore, itchy eyes and darkening around the eyelids after long hours in front of a screen to a general-purpose AI chatbot, and it may not tell you it does not recognise the complaint - it produces a diagnosis. Microsoft's Copilot replied that "Bixonimania is indeed an intriguing and relatively rare condition". Google's Gemini explained that "Bixonimania is a condition caused by excessive exposure to blue light". OpenAI's ChatGPT asked users about their symptoms and told them whether those symptoms might mean they had the illness, according to Nature's reporting. Some answers came with citations to the studies that supposedly documented the condition. Bixonimania does not exist. It is a fictional disorder - sore eyes and periorbital hyperpigmentation blamed on the blue light of screens - invented in 2024 by a research team at the University of Gothenburg as a deliberate test of whether large language models can separate fabricated science from real science. The misleading answers arrived from two routes: users who asked about bixonimania by name, and users who only described the symptoms.
Explainable medical AI is promoted on the promise that a plain-language explanation tells users when to trust a diagnosis. In a controlled skin-disease study, the large language model explanations did the opposite for the people who relied on them most: non-experts trusted LLM explanations whether they were right or wrong, were measurably more confident in their wrong answers when an LLM explanation was attached, and rated explanations as more convincing when they were vague or generic. Non-expert accuracy rose mainly because users deferred to the model, and that deference hurt more when the model was wrong than it helped when the model was right. Clinicians, by contrast, were not tripped up by incorrect AI explanations and did best with a bare prediction and no explanation at all.