+ Reply to Thread
Page 5 of 5 FirstFirst 1 5
Results 81 to 87 of 87

Thread: Unusual and Bizarre Uses and Behaviors of AI

  1. Link to Post #81
    Canada Avalon Member Johnnycomelately's Avatar
    Join Date
    14th January 2022
    Location
    Edmonton, Alberta, Canada
    Language
    English
    Age
    67
    Posts
    1,933
    Thanks
    25,178
    Thanked 12,006 times in 1,906 posts

    Default Re: Unusual and Bizarre Uses and Behaviors of AI

    This is presented as a feel good story, since he is said to have communicated the vulnerability to the robot vacuum maker, Chinese company DJI. DJI has said that they have since fixed it.

    I guess we are safe now, from having our homes cam’d and mic’d and map’d by the CPPCC.

    Crazy exploit. Am guessing it won’t be fully plugged, either by DJI’s stated action likely being a lie, or because other various smarter evil-worker nerds can access it. Smart Homes are compromised homes.

    I’m of half a mind to subscribe to some A.I. I realize that that may present temptations to my ego, but I feel confident in my growing ability to discern and dissent. I’ve always loved using tools, and these ones seem intruiging. Wish me luck? — not yet. I’ll tell y’all when I take that baby step.

    https://www.popsci.com/technology/robot-vacuum-army/

    Man accidentally gains control of 7,000 robot vacuums

    Sammy Azdoufal just wanted to steer his DJI Romo with a gaming controller.

    MACK DEGEURIN
    PUBLISHED FEB 21, 2026


    Quote A software engineer’s earnest effort to steer his new DJI robot vacuum with a video game controller inadvertently granted him a sneak peak into thousands of people’s homes.

    While building his own remote-control app, Sammy Azdoufal reportedly used an AI coding assistant to help reverse-engineer how the robot communicated with DJI’s remote cloud servers. But he soon discovered that the same credentials that allowed him to see and control his own device also provided access to live camera feeds, microphone audio, maps, and status data from nearly 7,000 other vacuums across 24 countries. The backend security bug effectively exposed an army of internet-connected robots that, in the wrong hands, could have turned into surveillance tools, all without their owners ever knowing.

  2. The Following 9 Users Say Thank You to Johnnycomelately For This Post:

    Bill Ryan (1st March 2026), Ewan (1st March 2026), grapevine (1st March 2026), Harmony (5th April 2026), kudzy (1st March 2026), sdv (13th April 2026), Snoweagle (19th March 2026), tessie999 (1st March 2026), Yoda (17th March 2026)

  3. Link to Post #82
    UK Avalon Founder Bill Ryan's Avatar
    Join Date
    7th February 2010
    Location
    Ecuador
    Posts
    39,288
    Thanks
    285,778
    Thanked 521,776 times in 37,823 posts

    Default Re: Unusual and Bizarre Uses and Behaviors of AI

    At the start of this new video Scott Ritter explains how it was Anthropic's AI program Claude that told the US military that the school in Minab that was destroyed by Tomahawk missiles was part of an Iranian IRGC base.

    No human checked the AI information. 170 girls died.


  4. The Following 13 Users Say Thank You to Bill Ryan For This Post:

    Ewan (17th March 2026), grapevine (17th March 2026), Harmony (4th April 2026), Inversion (17th March 2026), Johnnycomelately (18th March 2026), kudzy (17th March 2026), rgray222 (20th March 2026), Rizotto (17th March 2026), sdv (13th April 2026), Snoweagle (19th March 2026), Tintin (Yesterday), Vicus (17th March 2026), Yoda (17th March 2026)

  5. Link to Post #83
    Avalon Member rgray222's Avatar
    Join Date
    24th September 2010
    Language
    English
    Posts
    3,401
    Thanks
    13,825
    Thanked 31,656 times in 3,290 posts

    Default Re: Unusual and Bizarre Uses and Behaviors of AI

    Incredibly easy, incredibly deceptive and incredibly convincing - not sure how we can trust video and image information from now on. (49 second video)


    Last edited by rgray222; 5th April 2026 at 00:11.

  6. The Following 8 Users Say Thank You to rgray222 For This Post:

    arjunaloka_official (12th April 2026), Bill Ryan (12th April 2026), Ewan (5th April 2026), Harmony (4th April 2026), Johan (Keyholder) (4th April 2026), Johnnycomelately (12th April 2026), Tintin (Yesterday), Yoda (5th April 2026)

  7. Link to Post #84
    Canada Avalon Member Johnnycomelately's Avatar
    Join Date
    14th January 2022
    Location
    Edmonton, Alberta, Canada
    Language
    English
    Age
    67
    Posts
    1,933
    Thanks
    25,178
    Thanked 12,006 times in 1,906 posts

    Default Re: Unusual and Bizarre Uses and Behaviors of AI

    Sticky Trap for miscreant wannabe lawyers: a feel-good story about misuse of AI.

    https://edmontonjournal.com/news/loc...ety-suspension

    Edmonton articling student rebuked for using AI to draft appeal of law society suspension

    "The misuse of AI to write his submissions, and in particular his reliance on a hallucinated case in his submissions, suggests otherwise"

    Author of the article:By Jonny Wakefield
    Published Apr 11, 2026
    Last updated 9 hours ago


    Quote An Edmonton student-at-law is in hot water after Alberta’s legal regulator found he used artificial intelligence to draft an appeal of an earlier suspension.

    Manraj Tiwana was suspended for 20 months last year after the Law Society of Alberta found he was taking clients without telling his articling supervisor.

    When Tiwana appealed the decision, he found himself facing an even sterner rebuke after admitting he drafted his submissions using AI, which cited a “hallucinated” case that did not actually exist.

    “While we appreciate that Mr. Tiwana was candid in admitting his use of AI, he did not seem to grasp the gravity of misusing it in this way, particularly in this setting, where he is arguing that his discipline was too harsh and his prospect for recovery is good,” the seven-member appeal panel wrote in its March decision.

    The misuse of AI to write his submissions, and in particular his reliance on a hallucinated case in his submissions, suggests otherwise.”

    ‘Abject dishonesty’

    In 2022, the law society received complaints that Tiwana had been taking off-the-books clients unbeknownst to the lawyer serving as his principal. Tiwana accepted payment from the clients personally, without telling them his principal was being kept in the dark about the arrangement.

    The law society investigated but was delayed “because Mr. Tiwana repeatedly misled the investigators,” the appeal panel wrote.

    In 2025, the then-35-year-old pleaded guilty to six disciplinary charges and agreed to a 20-month suspension — a punishment the hearing panel said was “as lengthy as it gets short of disbarment.” It also ordered him to pay $18,000 in costs. While Tiwana initially agreed to the suspension — which a seasoned defence lawyer hashed out with the law society — he later appealed, arguing it was too harsh. This time, Tiwana represented himself.

    According to the decision, a lawyer for the law society contacted Tiwana ahead of the appeal because she could not find a case cited in his submissions. Tiwana responded it was a “placeholder.” It turned out the case had been made up by an unspecified AI tool.

    The panel ultimately rejected Tiwana’s appeal, finding no legal problems with his earlier guilty plea. The panel reasserted a 20-month suspension was appropriate for someone who displayed “abject dishonesty” and made “long-running” attempts to deceive investigators.

    The panel also gave little credence to Tiwana’s claim he didn’t understand what a 20-month suspension meant.

    “He seems to suggest he was not aware, at the time of his plea, that when the 20 months was served he would have to apply for admission or reinstatement, and the result of that application was uncertain,” the panel wrote. “He also suggests he was not aware that, in that process, his credentials may have by then expired.”

    “These submissions, when made by a party represented before the hearing committee by experienced defence counsel, are difficult to accept when there is no allegation or evidence that his counsel was ineffective or failed to advise him.”


    ‘Haphazard approach’

    Tiwana’s use of AI was not the only problem during the appeal hearing.

    He also repeatedly referred to a Law Society of B.C. case “which had no relevance at all to this matter” and did not support any of positions he claimed it did, the panel said.

    When asked for the case’s citation and “style of cause” — the legal name of the case — “Mr. Tiwana struggled to understand what that meant,” the panel added.

    The panel ultimately decided Tiwana should pay full costs of the appeal hearing.

    “In considering what level of indemnity is required, and what costs burden would be fair, we find the most important factor in this case is Mr. Tiwana’s conduct of the appeal, including his haphazard approach to attempting to adduce new evidence, and, more troubling, his conduct in briefing and arguing the appeal,” the panel wrote.

    Attempts to reach Tiwana, who is not listed in the provincial lawyer directory, were not successful.
    Last edited by Johnnycomelately; 12th April 2026 at 09:01. Reason: Add link to article

  8. The Following 5 Users Say Thank You to Johnnycomelately For This Post:

    Bill Ryan (12th April 2026), Ewan (12th April 2026), Harmony (12th April 2026), kudzy (12th April 2026), Yoda (12th April 2026)

  9. Link to Post #85
    Germany Avalon Member arjunaloka_official's Avatar
    Join Date
    25th July 2023
    Language
    German
    Posts
    120
    Thanks
    1,098
    Thanked 696 times in 105 posts

    Default Re: Unusual and Bizarre Uses and Behaviors of AI

    Quote Posted by rgray222 (here)
    Incredibly easy, incredibly deceptive and incredibly convincing - not sure how we can trust video and image information from now on. (49 second video)
    Short answer: by using intuition.
    Disclaimer: The above is only a mystical hypothesis in form of a question, but neither factual statement, nor request, nor advice.

  10. The Following 5 Users Say Thank You to arjunaloka_official For This Post:

    Bill Ryan (12th April 2026), Harmony (12th April 2026), Johnnycomelately (13th April 2026), rgray222 (18th April 2026), Yoda (18th April 2026)

  11. Link to Post #86
    Avalon Member sdv's Avatar
    Join Date
    5th March 2012
    Location
    On a farm in the Klein Karoo
    Posts
    1,303
    Thanks
    5,241
    Thanked 5,781 times in 1,176 posts

    Default Re: Unusual and Bizarre Uses and Behaviors of AI

    Quote Posted by Bill Ryan (here)
    At the start of this new video Scott Ritter explains how it was Anthropic's AI program Claude that told the US military that the school in Minab that was destroyed by Tomahawk missiles was part of an Iranian IRGC base.

    No human checked the AI information. 170 girls died.

    It was not only the school that was bombed, with a double tap strike (waiting for rescuers to get there to rescue the children, most of whom were still alive when the second missile was dropped). On that site, a clinic serving the local population and a storage building for medical supplies was also destroyed. I understand that Americans do not want to accept that this could have been deliberate and so are blaming AI, but the evidence shows that the bombing of the school was part of a pattern and not an AI error. Research what else was targeted that day, and in all the subsequent days. You don't destroy a civilization by targeting its military (which posed no threat to the US) ... you bomb schools (there were others), recycling plants for domestic waste, historical buildings used as museums and cultural centres, a sports centre for children, clinics (lots of them), emergency response centres, civilian apartment blocks, parks where children play ... and then bridges, etc.

    If AI chose those targets, it was programmed to do so, and that programming was done by humans.
    Sandie
    Somewhere, something incredible is waiting to be known. (Carl Sagan)

  12. The Following 8 Users Say Thank You to sdv For This Post:

    Bill Ryan (18th April 2026), bojancan (13th April 2026), Ewan (13th April 2026), ExomatrixTV (18th April 2026), Harmony (14th April 2026), Johnnycomelately (14th April 2026), kudzy (13th April 2026), Yoda (18th April 2026)

  13. Link to Post #87
    Netherlands Avalon Member ExomatrixTV's Avatar
    Join Date
    23rd September 2011
    Location
    Netherlands
    Language
    English, Dutch, German, Limburgs
    Age
    60
    Posts
    29,661
    Thanks
    44,341
    Thanked 166,089 times in 27,686 posts

    Default Re: Unusual and Bizarre Uses and Behaviors of AI

    • Anthropic Banned My Claude Account:

    Anthropic banned my Claude account with no warning, no explanation, and no support. Why? I have no idea, and I probably never will, but I do have some thoughts on free and fair access to AI... Chapters: 0:00 - Claude Suspended My Account 4:32 - Is AI a Human Right? 7:29 - Competition and Open Source 9:36 - Can Frontier Labs Survive Themselves ?
    • @alpha-terry quote: Got banned today as well. You hit the nail on the head, I had all the same thoughts. Wondering how, as a developer, I'm going to possibly compete without something like claude code. Feeling like a second class citizen. Of course there are ways around this, ways against their TOS, but I'm sure it can be done if necessary. Hoping there is enough noise to get them to change their policy.
    • @anderstaylor3377 quote: Yeah, I got banned in precisely this same way. Absolutely infuriating.
    • @scottlong7319 quote: Imagine if you were about to discover something and Claude shut you down to steal it or prevent you from discovering it.
    • @codegoXn quote: The same thing happened to me, I don't know what to do now
    • @HitregStudios quote: Legit same thing happened to me. No warning, I can't find a violation that I even came close to in their usage policy. I feel completely neutered here. The only thing I could guess was during their token stealth nerf at the end of March was when this happened to me... Maybe when that issue was present caused something???
    • @JoeThompson-t4l quote: This exact thing happened to me too! Not a single questionable use and only had it for 48 hours. Best guess I have is that I use a VPN?
    • @HitregStudios quote: Legit same thing happened to me. No warning, I can't find a violation that I even came close to in their usage policy. I feel completely neutered here. The only thing I could guess was during their token stealth nerf at the end of March was when this happened to me... Maybe when that issue was present caused something???
    • @fnapresents quote: Just got my ban this morning - I think it’s because my daughter was in the room talking while I was using Claude voice. All my work is just gone…
    • @rjjarecki quote: Got a very similar email from OpenAI, they presented a follow up window, but declined my follow up request. They would not tell me what channel the error occurred on, if someone has stolen my api Keys and is 'commiting fraud' in a discrete manner, or if their system is just hallucinating something.
    • @lHARDlCaps quote: I'm trying to use AI to enforce the laws that protect the human rights of disabled Medicaid recipients. The way Anthropic operates, preventing third party harnesses from using Claude MAX, trying to lock people into their ecosystem, then banning people without any fair hearing or appeals process, it worries me a lot. Frankly, I think Anthropic is failing as a public good company. You can't work in the public good when you don't have effective communication and support with the people working in the public good. On top of that, anthropic has policy and practices that undermine work like mine. I think things could easily blow up in Anthropics face. They seem to have a total disregard for how they negatively affect users like me and all it's going to take is one ban at the wrong time. It's like Anthropic is playing Russian roulette.
    • @Stewart-x8k quote: The same exact thing happened to me. Am I going to be charged next month???
    • @GordyP7 quote: Yeah, I also got suspended yesterday and I have no idea why. I got an invoice automatically for a refund, and it initially showed up in my Wells Fargo account, but now it has randomly disappeared, and it hasn't been refunded. What do I do about this? If I get the refund okay ill go to a different AI code, but I haven't been refunded.
    (back up video in case it gets censored):

    (back up video in case it gets censored)
    Last edited by ExomatrixTV; 18th April 2026 at 20:26.
    No need to follow anyone, only consider broadening (y)our horizon of possibilities ...

  14. The Following 9 Users Say Thank You to ExomatrixTV For This Post:

    Bill Ryan (18th April 2026), Ewan (18th April 2026), Harmony (18th April 2026), Johnnycomelately (19th April 2026), petra (18th April 2026), rgray222 (18th April 2026), Sunny (18th April 2026), Tintin (Yesterday), Yoda (18th April 2026)

+ Reply to Thread
Page 5 of 5 FirstFirst 1 5

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts