If anyone needs an endless stream of examples why it's a bad idea to caption your images by carelessly adding an unedited LLM-generated image description without any kind of disclaimer, @HourlyPornhubbedHeathcliff is a GREAT example of this wildly unhelpful practice!

Pick literally any of them I'm sure (and be aware they are captioned with pornhub comments and thus likely to contain NSFW text), but this one just came across my feed and the hallucinations are particularly egregious: https://masto.ai/@HourlyPornhubbedHeathcliff/114813848311864253

Oddly enough, while that particular image/comment is totally SFW, the LLM hallucinated a mildly-NSFW phrase into the description, I'm guessing generated from the scrawled signature.

All this might be hilarious, if it wasn't leaving visually-impaired people frustrated and with very little clue as to what the hell is actually happening, and illustrating what is undoubtedly happening across the web in much more important contexts than silly comic rewrites.

#Accessibility #GenerativeAI #PlausibleSentenceGenerators #ImageDescriptions #LLMs #a11y #BadAccessibility

Hourly Pornhubbed Heathcliff 🔞 (@[email protected])

Attached: 1 image #Heathcliff #NSFW #bot #lewd

Mastodon

Microsoft - like every corporation - is insatiably horny for firing workers. It spent the past three years cutting its writing staff to the bone, with the express intent of having AI fill its pages, with humans relegated to skimming the output of the #PlausibleSentenceGenerators and clicking "OK":

https://www.businessinsider.com/microsoft-news-cuts-dozens-of-staffers-in-shift-to-ai-2020-5

We know about the howlers and the clunkers that Microsoft published, but what about all the other travel articles that don't contain any (obvious) mistakes?
15/

Microsoft News cuts dozens of workers in shift to AI-driven news

The news aggregation giant is leaning more on artificial intelligence and away from human editors for MSN.com.

Insider