back to article If your AI does the crime, you'll do the time, warns DoJ

If juggling the extreme cost and hazy ROI of AI weren't enough of a headache, the United States Department of Justice (DoJ) now expects enterprise compliance officers to start weighing the tech's potential for harm – or risk stiff fines if it breaks the law. Nicole Argentieri, the principal deputy assistant attorney general …

  1. Homo.Sapien.Floridanus

    DOJ: Your company, devoid of humans, has been run amok by AI whose bots have broken a plethora of laws ranging from money laundering to copyright infringement and data theft. Do you wish to have a lawyer present before answering our questions?

    CEO: Siri, activate lawyerbot app.

    1. elsergiovolador Silver badge

      Siri: Activating your laundry. Standard program set.

    2. Jedit Silver badge

      "AI whose bots have broken a plethora of laws"

      Sorry to piggyback on your joke with a serious comment, but: pretty much every AI out there has been solely trained on data taken without express permission and often in infringement of copyright. Applied as described, this law will result in the prosecution of every company touting an AI. Which is all to the good, of course - the sooner we eradicate this bullshit, the better.

  2. commiepinko

    Artificial Xenophobia

    The resumé screening apps used by a large majority of US employers are trained on public data sets. Consequently, they are demonstrably and quantifiably sexist, racist, homophobic, Anti-Semitic, more, and worse. They are the codification of some of humanity's most destructive faults, and they are the first barrier anyone who wants a grownup job must cross. They're the most energetic, competent, relentless bigots the world has ever seen, committing the most heinous sorts of discrimination with a thoroughness and accuracy that's the stuff of a Klansman's wet dreams.

    And every company using them is in direct, flagrant violation of a host of U.S. federal and state fair employment statutes. The speed, scale, and stealth of the assault is as devastating as it is breathtaking.

    If ignorance of the law is no excuse, neither is ignorance regarding your AI. Every last one of those companies should be prosecuted to the fullest extent of the law, and the perpetrators behind the production of the software given prison terms. The production of some software must be regulated and monitored, and the costs of doing so must be paid by the producers. "To encourage the others."

    1. Bendacious Silver badge

      Re: Artificial Xenophobia

      I’d upvote twice if I could. It’s a terrifying thought that lives could be blighted by algorithms that you have no idea even exist. Decisions made about your life by an algorithm that cannot explain how it came to a decision. There are existing laws to prevent bias in a lot of situations but you have to know it’s happening. It could be years or never that it comes out. Governments and ML companies know it is coming but I don’t have much faith they can slow the money machine enough to fix it.

      1. veti Silver badge

        Re: Artificial Xenophobia

        Humans have been discriminating against one another since time began, and for nearly all of that history they were just as unaccountable as any AI. Eventually, people figured out how to construct experiments to test and statistically demonstrate the existence and extent of discrimination. There's no reason those same experiments shouldn't still work just as well today as they did in the 70s.

        1. MacGuffin

          Re: Artificial Xenophobia

          Is there a reason the SHOULD work just as well?

        2. Phil Koenig

          Re: Artificial Xenophobia

          "veti" wrote:

          and for nearly all of that history they were just as unaccountable as any AI.

          Not really.

          Totalitarian societies are gonna be totalitarian.

          But in a supposedly open, democratic society, if the government or a private organization makes a stupid decision about something we can subpoena the individuals involved and question them about it, sanction them for it, throw the bums out (or even in jail) and change the laws that enabled them in the first place if necessary.

          None of that would have any impact on an "AI" which cannot even tell us why it made a particular choice.

          Ergo, those who put such an "AI" in a position to make important decisions are the ones that have to answer for the AI's crimes.

          For once, the US Government is showing signs of actually understanding the risks of new technology and putting in some reasonable controls before it has laid waste to the entire country.

          1. M.V. Lipvig Silver badge

            Re: Artificial Xenophobia

            Yup. My own company now uses an AI interview to hire and promote. This is the first company where my face, and not my job performance, decides whether or not I get promoted. Every other company I've ever worked at, I've made senior tech within 5 years. This one, been here 6 years and I'm still a tech 1.

            To be fair (to the company) there was a hiring and promotion freeze for 4 years, but when they opened it up they had AI going. To be unfair, I work for a company with a heavy DEI presence, and I'm a white male.

            I'm only not concerned because I've already set myself up for retirement and am planning to leave in a couple of years anyway, and am really just working for benefits and to finish out a few projects for my current house before I sell it. I don't want to pull any investments for this.

    2. Anonymous Coward
      Anonymous Coward

      Re: Artificial Xenophobia

      Xenophobia is an overloaded term often hijacked and used against liberal countries.

      Having lived in many such countries, cultural and ethnical clustering was obvious. You can not force it with a policy. It is not working and never will. Mixing cultures takes generations. Mixing religions - millennia.

    3. Anonymous Coward
      Anonymous Coward

      Re: Artificial Xenophobia

      > codification of some of humanity's most destructive faults

      People must take decisions and have no time to experiment for causality, therefore they use ethnical correlations as the 1st filter.

      Personal reputation has always been the best alternative. But somehow the liberals are against personal scores. Why should it be difficult to check backgrounds? Especially, if an employee is hard to fire because of socialist laws, while obviously no employer wants a lazy irresponsible person with criminal tendencies. How such economy can compete with other countries?

      Above is just one example of counterproductive philosophy leading to the "xenophobic" but working shortcuts in human relations.

      1. Anonymous Coward
        Anonymous Coward

        Re: Artificial Xenophobia

        > Personal reputation

        Institutions play the key role by "issuing certificates of reputation". This could be one single reason why some countries succeed, while others fail. Corruption is basically devaluation of such certificates.

      2. veti Silver badge

        Re: Artificial Xenophobia

        If you're such a big believer in "personal reputation", why are you posting anon?

        I don't know, but I'm guessing the broken English in your last two sentences, and the pejorative generalisations about "the liberals" and (unspecified) "socialist laws" might give clues.

        Why should it be difficult to check backgrounds? - news to me that it is, perhaps you'd like to give examples?

        1. Anonymous Coward
          Anonymous Coward

          Re: Artificial Xenophobia

          > why are you posting anon

          You are an anon too. There are several valid reasons I post double-anon.

          > broken English

          - this is a fresh example of "xenophobic" judgement shortcut just from a few sentences posted by someone online. I guess you would not have mentioned my broken English if we met face-to-face by hiding your xenophobic thoughts.

          > give examples

          - En el Reg comments I proposed to introduce "antisocial" scores for shop lifting and other petty crime. And got downvoted badly.

      3. LBJsPNS

        Re: Artificial Xenophobia

        "the liberals are against"

        Stopped reading there.

        1. Anonymous Coward
          Anonymous Coward

          Re: Artificial Xenophobia

          So you stopped reading. Very interesting. Any other thoughts about the post overall? People come here to share ideas.

          Yours is another example of thought shortcuts in human relations. Communication can stop from just one incorrectly used word. Or even may cause a conflict.

      4. Phil Koenig

        Re: Artificial Xenophobia

        Anonymous Cowardly Poster wrote:

        People must take decisions and have no time to experiment for causality, therefore they use ethnical correlations as the 1st filter.

        Translation:

        "I am a bigot, so therefore all must be bigots"

        I never fail to be amazed at the startling lack of self-knowledge so many people exhibit.

        Are you an AI?

    4. amanfromMars 1 Silver badge

      For revelations and deeper understandings of obscure and obfuscating programs and projects.

      Re: Artificial Xenophobia ..... Nice first "publicly visible post • joined 26 Sep 2024", commiepinko. Welcome to El Reg, biting the hand that feeds IT publishing situations worthy of peer review and independent prime premium forensic scrutiny/reverse engineering dissection.

  3. spold Silver badge
    Holmes

    Well...

    ...since the AI just replaced your lawyers this is rather moot...

  4. Dagg Silver badge
    Joke

    Q. What is a 1,000 AIs at the bottom of Sydney Harbour?

    A. A good start!

  5. Pascal Monett Silver badge

    Good

    Since actual AI does not exist, making sure companies are responsible for whatever their statistical analysis machine spews is a good decision.

  6. Anonymous Coward
    Anonymous Coward

    Two tier justice

    I bet the state is not subject to the same justice when their AI breaks the law.

    1. MacGuffin

      Re: Two tier justice

      US President’s AI is already fullly immune. As SCOTUS.

  7. Anonymous Coward
    Anonymous Coward

    I wonder if this will also extend to road traffic accidents when Musk's Autobots are at fault

    1. John Brown (no body) Silver badge

      Tesla Autopilot and the Waymo crashes/deaths were probably part of the evidence looked at that lead to this.

  8. Philo T Farnsworth Bronze badge

    Just remember. . .

    You can't spell JAIL without AI.

    1. ZX8301

      Re: Just remember. . .

      Gaol

  9. azander

    Treat the AI like an employee....

    If the DOJ was serious about cracking down, they should treat the AI as an Employee of the company.

    However since it is not human or a citizen of the US, and is a physical Asset like any other computer or program, it can't claim any special status, plead for immunity or even use the 5th Amendment to protect itself from self incrimination. If the DOJ does that, then the companies would see the AI as a very large liability instead of something good for business. It would do what they want, with very little cost to the DOJ.

    DOJ: Here is a warrant for the collection and incarceration of your AI. We will be questioning it at our leisure. No, you can't have an attorney present, it has no rights.

    Company: We'll sue!

    DOJ: Of course you will. Too bad your AI is an "Asset", not an employee, that can be searched and queried without representation or advocacy.

    Company: Bankruptcy!

    1. GenericLeftieWhackjob

      Re: Treat the AI like an employee....

      Easy, but that assumes people go into government to actually do the job of governing and not to amass power and influence.

    2. An_Old_Dog Silver badge

      Bankruptcy

      Pilot (over aircraft interphone) Eject! Eject!

      Tower: I see ... five good golden chutes!

      Pilot: (hanging from his chute): Banner away!

      (Towed behind aircraft - a banner): "BANKRUPTCY"

  10. Phil Koenig

    Competent Governmentz, I Gotz It

    For once the US Government shows signs of actually understanding the impacts - including potential negative impacts - of emerging technology. BEFORE it lays waste to the countryside.

    You can say goodbye to all of that - in fact a literal weaponization of corporate bad behaviour - if the former Con Artist in Chief manages to get back in.

    How pathetic that we are even faced with such a possibility. Again.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like