Thank you for that analysis.
I understand why you say that but I have to say I’m not so sure of that last part - juries can be fickle and it could be reasonably argued that the dead user had no reason to doubt the AI and was not au fait with how wrong it could be - they were not the sort of user who paid attention to AI news stories and just used Google as a trusted source. A jury of similarly unsophisticated / tech-innocent users might possibly agree with that argument.
But IANTL, of course. 
When I was in school, they put the fear of God into us about eating wild mushrooms, (also just the fear of God in general).
I wouldn’t dream of eating a wild mushroom without an expert in mushroomology.
Yeah, maybe a jury could reach that conclusion. But this statement is at the bottom of the Google AI summaries right now: “Generative AI is experimental.”
Bottom line…if anyone you know trusts any AI to give them accurate information, please disabuse them of that notion. Immediately. Even if Google were found liable, it ain’t bringing your loved one back.
Agree . I eat several dozen wild plants. I eat one wild mushroom ( giant puffball). Fungi are trickier than plant identification.
This, to me, is at the heart of why I roll my eyes when I hear people foaming about AI. A truly intelligent entity asks questions when they have incomplete information. AI just confidently strides forward with whatever shit floats to the top.
Ai info seems to be an aggregation of “internet wisdom” and we know that a large group of human’s posting stuff to facebook is Always correct. So what could be dangerous?
“Google AI just Might Kill You.” But nah. Look, if you’ve reached a point in your life where you are eating a wild mushroom based on Google Search results, you’ve made a long series of poor life choices on your way to the terminal.
Ah - them’s the words that’ll do it. “Ladies and gentlemen of the jury, the plaintiff was clearly, and without doubt knowingly, indulging in their own experiment, and they found out what its results were.”
Holy shit - from the comments on that post, another case a few months later with more harrowing - and expensive! 
- results (comment Jan 2007):
A day or two before 9 September 2006, Tim Boerge consumed pale Amanita phalloides in Central California which he and his wife had spore printed and concluded that the spore print was pink (due to staining of the paper, rather than a true spore print) – but he only took the spore print and tried to identify the mushrooms after he had eaten them and started to feel ill – sound a bit familiar? His poisoning was really serious – after 30 delirious hours at a hospital in Watsonville where he was in IVs for dehydration and lots of oral charcoal, he was transported to UCSF where he suffered for 3-4 days and wound up with a liver transplant. For three full weeks he was in incredible pain and thought he would die. Five months later he is still weak and in some pain. His medicines are currently running $1500/month. The Watsonville hospital stay cost over $150,000, the ambulance ride to San Francisco cost $34,000 and the transplant tab pushed his total costs to over $800,000 – and he has no insurance. Amazingly he is now simply happy to be alive.
Not planning on foraging wild mushrooms, but thankful that the NHS is there if I were to do so - rather than have to sell my house to pay the bills 
Does anyone know what the state of genetic identification is for mushrooms? Or immunoassays for toxins of concern?
Obviously a full genome sequence isn’t going to be financially viable; but are the mushrooms of interest well characterized enough(and do they possess sufficient exploitable markers) that significantly smaller sequences can be used to get a good ID on a sample?
That approach obviously lacks the immediate gratification and minimal cost of machine vision; but when the fuck around/find out loop potentially involves mycotoxins the downside risk is alarming enough that at least the cheaper flavors of more robust identification techniques start to look interesting if ‘just maybe don’t eat it’ isn’t your preference.
No!
“Older” is not better in case of fungi. I have several books which were really good at the time, and are still great for identifying most mushrooms. But never, ever rely on those old books in regard to edibility. I repeat: do not trust older books on edibility of mushrooms.
There is a number of species in my books which were considered edible at the time and it turned out that they can be harmful. I had one case - sadly can’t remember what it was - where recent research suggests that about a third of the population would get an immune reaction which is likely destructive to your kidneys if (NB!) they had been exposed to that specific mushroom before.
My family used to go foraging until the Chernobyl disaster, and mushrooms were an integral part of my upbringing. They also used to quip that ‘everything is edible, but sometimes just once’. With mushrooms, according to science, it might kill you the second time.
Short: only trust the ones you are really, really, really sure about. (And even then, do not pair them with alcohol. Just… don’t.)
ETA: @RickMycroft, do I owe you a coke or doesn’t it count when my family was quoting Pratchett before he wrote this?
But after that little mishap mushrooms were so much easier to find if you had the right equipment!
Fair enough. Times and best info changes. I keep my foraged edibles contained to chicken and hen of the woods, morels, chanterelles, lion’s mane, and the oyster fam. All easy to identify, and all delicious.
But yes, be sure before you eat, cook before eating, stop eating if you have a reaction and don’t pair with alcohol. If in doubt, don’t.
Gemini: “Destroying Angel is fine if you cook it with glue.”
This topic was automatically closed after 5 days. New replies are no longer allowed.
