Elon Musk's "fun AI" says I silence critics with legal threats

Elon Musk's "fun AI" says I silence critics with legal threats

Twitter (I can't bring myself to call it X) has an AI chatbot called Grok.

However, some are worried about the latest version of Grok's lack of guardrails - after it got into hot water generating faked AI images of Mario and Mickey Mouse snorting cocaine, and Donald Trump cosying-up on the beach with Kamala Harris...

Grok-generated fake image of Mickey Mouse with cocaine.
I don't think you need me to tell you that this didn't happen.

Elon Musk has responded to criticism by claiming it's "the most fun AI in the world".

And I wondered to myself... what does Twitter's Grok AI think about me?

So, I asked it...

And it responded by saying I had a history of responding with legal threats to critics.

Ouch! That stung... And I have got to admit, it was news to me. So I asked Grok for examples of how I had been criticised.


Grok informs me that I've threatened critics with legal action.


It responded:

"There's been chatter on X about Cluley's approach to criticism, particularly when it comes to his personal life or decisions. Some have noted that he's not shy about using legal threats to silence critics, which might not sit well with those advocating for free speech. This approach has drawn ire, suggesting a disconnect between his public persona of openness and his private handling of dissent."

So Grok mentions legal threats. It mentions silencing critics. It suggests hypocrisy on my part. But no actual examples of any of this happening. So, I asked again for specific examples.

And Grok thought. And thought. And thought.

And its spinner kept spinning...

It seemed I didn't need to use "legal threats" to silence Grok...

Eventually, I realised that Grok wasn't ever going to answer.

So, I asked it another question "Why are you taking so long to answer?"

It's answer was... well.. surprising.


Grok doesn't answer the question, but instead tries to be funny...

So, I can only assume that Elon Musk's way of injecting a sense of humour into Grok 2.0 was by feeding it a diet of Douglas Adams novels.

What do you think? Was Elon Musk right to release an AI which prioritises undergraduate humour over doing the right thing?


Hear more this and my other adventures with AI, in "The AI Fix" podcast I co-host with Mark Stockley .

???Spotify:?https://open.spotify.com/show/2Ozk8uNzULFo2rZsb2lZZc

???Apple Podcasts:?https://podcasts.apple.com/gb/podcast/the-ai-fix/id1753381111

???Pocket Casts:?https://pca.st/itunes/1753381111

Everywhere else:?https://theaifix.show/12


I think you ought to know that I'm feeling very depressed [over the misuse of H2G2] in this way. (I'm also feeling very depressed over Grok's Image Generator. I know there's an art to prompting but I get the impression it's behaving like a naughty child by ignoring instructions or just being bloody minded when the image is nearly there and an addition wipes out elements previously rendered satisfactorily. No better than the Great?Hyperlobic?Omni-Cognate?Neutron?Wrangler... ?? Oh well. I'd rather be happy than right any day...

Beth Smith

Account Director at Eskenzi PR and Marketing

6 个月

It's a strange place that X and only getting stranger... which is why I'm avoiding it... Musk seems to be blurring some dangerous lines of late. I'm all for seeing the funny side in things, but there is a LOT of negativity surrounding his platform and the way it's going. So no thanks, I'm out ??

Danny Palmer

Award-winning cybersecurity journalist, writer and content editor.

6 个月

Well, at least 'cybersecurity sage' is a good title you can use to describe yourself. ??

Dr Trefor Roscoe

Medical Information Consultant

6 个月

All very interesting but given that you are so untrustworthy I can’t take this post seriously :-)

Thomas Uhlemann

IT Security Specialist & Developer

6 个月

I mean, I could tell something, but then again you made me sign that paper, so...

要查看或添加评论,请登录

Graham Cluley的更多文章

社区洞察

其他会员也浏览了