Uh-oh: Artificial intelligence can now write a complex essay on an obscure academic topic in 10 seconds

17
890


A new “chatbot” created by artificial intelligence company OpenAI Inc. is roaring onto the internet scene after being in stealth mode, as robot intelligence has grown more sophisticated behind the scenes for years.

OpenAI released ChatGPT, a bot that converses easily with humans via artificial intelligence. The bot can help humans write computer code, write essays and poems, and strategize.

Here’s an example of an academic essay that ChatGPT produced within 10 seconds of being asked:

It’s not just high school composition teachers that should worry. The rise of artificial intelligence has grave possibilities for humanity, and may be as dangerous as nuclear war, some futurists are warning. The bots are so sophisticated that what humans interact with on social media may or may no longer be a real human.

The website, thispersondoesnotexist.com shows just how far the deception can go. Every time a user refreshes the page a new face is shown, completely made up by artificial intelligence but appearing 100% human and believable.

A chatbot can now create a persona, access Twitter or Facebook, and start convincing people that a political candidate or ballot measure is good, bad, or even criminal. The voters would never know they are being interacted with by a programmed bot.

A chatbot can now go into a Facebook group, (think Juneau Community Collective, Bethel, Spenard, or Save Anchorage) and create dozens of online personas. Then the chat bot could start conversations, and persuade people to behave in a certain way or believe something is going to happen.

Facebook has also recently unveiled an online AI bot called CICERO, which has reached “human-level performance” in the board game called Diplomacy. This is a game built on human interactions, rather than moves and maneuvers as you would find in chess or checkers. The bots learned about human behavior by playing against thousands of humans and learning the complexities of how humans approach the game, in which the goal is to take over the world.

The bots learned about ways that humans use strategy or act impetuously. They learned about when humans get fatigued and how that affects their decisions. At this point, the CICERO bot is winning all the time, and continues to learn about humans by playing the game with those who probably don’t realize they are training a chatbot.

What is OpenAI? It’s a research lab founded by Elon Musk and Sam Altman, among others, in 2015. The group collectively pledged $1 billion to the project. Musk has resigned from the board in 2018, citing “a potential future conflict (of interest)” with Tesla AI, which he was developing for self-driving cars. He remained a donor, and in 2019, OpenAI received a $1 billion investment from Microsoft, which is now the company’s official preferred vendor.

Scientists such as Stephen Hawking have expressed concern that if AI gains the ability to re-design itself at an ever-increasing rate, “an unstoppable “intelligence explosion” could lead to human extinction.”

Even Musk says that AI is humanity’s “biggest existential threat.”

The rise of these technologies points to the need in Congress and in the White House for people who understand how complex these artificial intelligence matters are.

For now, Alaska is represented by “Fish, Family, Freedom” Congresswoman Mary Peltola, who does not express the capacity to understand the threats posed by AI, nor the interest in resolving the ethical, moral, and legal issues. Whether Sen. Lisa Murkowski or Sen. Dan Sullivan have the capacity — or the courage — to take on this growing threat is yet to be seen.

17 COMMENTS

  1. Great article Suzanne. You are right! “The rise of these technologies points to the need in Congress and in the White House for people who understand how complex these artificial intelligence matters are.

    For now, Alaska is represented by “Fish, Family, Freedom” Congresswoman Mary Peltola, who does not express the capacity to understand the threats posed by AI, nor the interest in resolving the ethical, moral, and legal issues. Whether Sen. Lisa Murkowski or Sen. Dan Sullivan have the capacity — or the courage — to take on this growing threat is yet to be seen.”

    • Dark Moon, I see nothing but “positives” here! Why, all we’ll have to do in life is stay in bed and be our sinful selves and let the bots pray for our wasted souls. Where do you see the downside to this? And as for those poor jackasses who are incapable of analytic or creative thought and who can’t read, write, add, or subtract, and who walk around with blank faces and hollow eyes, why, they needn’t be bothered with any of the trivialities of their mundane existence: they’ll be able to spend their time guzzling away their drinks and listening to the counsel of higher intelligence!

  2. Ok, here’s my question for the bot:
    What category of mental afflictions best defines a Democrat, and what remedy or treatment is needed to resolve the disorder?

  3. There are a ton of incompetent folks in customer service that should be replaced by an AI. How can this be all bad?

  4. What can go wrong with artificial intelligence? We have a president and vice president of this country with no intelligence who mumble obscure messages on a daily basis (maybe not Kamala every day as she is often out finding “root causes” of problems). It’s all about controlling people remotely so no one has to take responsibility.

  5. Might be worth trying simple stuff first: unplugging the damned things, donating nearby real estate to China for EMP testing.
    .
    Instead of the plaintive: “can you write me a four paragraph academic essay…”, which doesn’t require artificial, or any, intelligence to answer with a simple yes or no, why not demand: “in 200 words or less, in German or French, in handwritten form, in one hour or less, with no electronic device, summarize the Anderson-Gellner theories of nationalism.”?
    .
    Then there’s all sorts of Turing Test variations which help intelligent human beings (and professors) figure out who or what’s communicating with them.
    .
    Who’s to say that work to confound AI inducing it to self-destruct isn’t just as advanced? In other words, must we believe AI is immune from the best, most determined efforts of the world’s black-hat hacking industry?
    .
    Just because it’s “AI”, we cower in fear, resigned to inevitable doom, to losing our stuff? If one can chum the political waters to induce Democrats to eat their own, what persuades us that the seeds of logical confusion can’t be introduced into AI causing it to consume itself?
    .
    And the Frankensteins who created this monster, who arguably have the key(s) to tame or destroy it, must we accept that they’re immune from the less-than-loving attentions of peasant masses whose lives, and livelihoods, though wrecked by AI, are still expected to provide necessities and comforts of privileged life to these masters?
    .
    Good news is AI, just like guns, bombs, diseases, pogroms, hostile rulers, can’t eliminate peoples’ will to survive.
    .
    Only people can make that choice.

  6. There is no intelligence without God. I suspect these machines are simply clever fast typing interns or otherwise communicating with spirits and not humans

  7. It would be interesting to see what AI has to say to about energy policy, and have it speak to the potential for a minimally powered if not totally powerless future due to catastrophic circumstances…

  8. Peltola is for family? Hasn’t she failed at marriage four or five times?

    Divorce is a terrible thing to do to kids.

    • Oh, Lisa O, you’ve made a good point! Run for congress: you’ve shown that you too are qualified! Who’d ever guess that English is your second language?

  9. Only a matter of time before we people invented a dummy who doesn’t require a ventriloquist and a hand to move its lips. We people talk to ourselves, ai is designed to hold a conversation picking up your body language,speech, emotions, and words. So. Talking to it is would still be
    talking to yourself. I could see the thing would be helpful to a campaign volunteer door knocking, when they need a partner.

Comments are closed.