Artificial Intelligence (AI) in General

[Dave from down under]

https://www.abc.net.au/news/2023-07-15/ ... /102591340

A Artificial intelligence experts have been asking each other a question lately: “What’s your p(doom)?”

It’s both a dark in-joke and potentially one of the most important questions facing humanity.

The “p” stands for probability.

The “doom” component is more subjective but it generally refers to a sophisticated and hostile AI, acting beyond human control.

So your p(doom), if you have one, is your best guess at the likelihood — expressed as a percentage — that AI ultimately turns on humanity, either of its own volition, or because it’s deployed against us.

The scenarios contemplated as part of that conversation are terrifying, if seemingly farfetched: among them, biological warfare, the sabotage of natural resources, and nuclear attacks.

These concerns aren’t coming from conspiracy theorists or sci-fi writers though.

Instead, there’s an emerging group of machine learning experts and industry leaders who are worried we’re building “misaligned” and potentially deceptive AI, thanks to the current training techniques.

They’re imagining an AI with a penchant for sleight of hand, adept at concealing any gap between human instructions and AI behaviour.

[Slim Cognito]

Life Imitates The Terminator.

[Foggy]

Why would AI be any less twisted than the Earthlings who built her? I keep telling y'all, you won't have real AI until you get the machines to have emotions, and then look out.

[neeneko]

Why would AI be any less twisted than the Earthlings who built her? I keep telling y'all, you won't have real AI until you get the machines to have emotions, and then look out.

I am more worried about the emotions of humans than AI. AI is a bit like 'god', you can make it tell you do whatever you want it to tell you, and then justify actions by saying 'the AI said so!'. Just look at how people are trying to use it in law enforcement.. 'we are not racist! the machine we trained with all our racist assumptions is telling us which people are dangerous!'

[RTH10260]

WormGPT: New AI Tool Allows Cybercriminals to Launch Sophisticated Cyber Attacks

THN
Jul 15, 2023

With generative artificial intelligence (AI) becoming all the rage these days, it's perhaps not surprising that the technology has been repurposed by malicious actors to their own advantage, enabling avenues for accelerated cybercrime.

According to findings from SlashNext, a new generative AI cybercrime tool called WormGPT has been advertised on underground forums as a way for adversaries to launch sophisticated phishing and business email compromise (BEC) attacks.

"This tool presents itself as a blackhat alternative to GPT models, designed specifically for malicious activities," security researcher Daniel Kelley said. "Cybercriminals can use such technology to automate the creation of highly convincing fake emails, personalized to the recipient, thus increasing the chances of success for the attack."

The author of the software has described it as the "biggest enemy of the well-known ChatGPT" that "lets you do all sorts of illegal stuff." It's said to use the open-source GPT-J language model developed by EleutherAI.

In the hands of a bad actor, tools like WormGPT could be a powerful weapon, especially as OpenAI ChatGPT and Google Bard are increasingly taking steps to combat the abuse of large language models (LLMs) to fabricate convincing phishing emails and generate malicious code.

"Bard's anti-abuse restrictors in the realm of cybersecurity are significantly lower compared to those of ChatGPT," Check Point said in a report this week. "Consequently, it is much easier to generate malicious content using Bard's capabilities."

Sophisticated Cyber Attacks

Earlier this February, the Israeli cybersecurity firm disclosed how cybercriminals are working around ChatGPT's restrictions by taking advantage of its API, not to mention trade stolen premium accounts and sell brute-force software to hack into ChatGPT accounts by using huge lists of email addresses and passwords.

The fact that WormGPT operates without any ethical boundaries underscores the threat posed by generative AI, even permitting novice cybercriminals to launch attacks swiftly and at scale without having the technical wherewithal to do so.



https://thehackernews.com/2023/07/wormg ... llows.html

[RTH10260]

Creator of Conservative Chatbot Powered by ChatGPT Says OpenAI Tried to Censor Content

By Bryan Jung
June 8, 2023Updated: June 8, 2023

The creators of a conservative chatbot powered by OpenAI’s ChatGPT announced that they shut down the chatbot because of pressure from OpenAI to censor the bot’s responses.

The conservative chatbot, named GIPPR in honor of the late President Ronald Reagan, debuted in May and is part of TUSK, a pro-free speech and anti-censorship web browser.

GIPPR is a modified version of OpenAI’s ChatGPT and provides users with answers from a conservative perspective.

The creators of GIPPR said they were forced to sever ties with OpenAI, after they were told that their chatbot failed to “conform to their requirements for what can or cannot be said.”

TUSK founder and CEO Jeff Bermant said in a statement to Fox Business that OpenAI told him that GIPPR was not in compliance with its policies, which were “specifically related to deceptive activity and coordinated inauthentic behavior” and that they needed to “keep users and third parties safe.”

The actions of OpenAI appear to many conservatives as another example of attempted censorship from the left.




https://www.theepochtimes.com/tech/crea ... nt-5320420

[RTH10260]

AI use in breast cancer screening as good as two radiologists, study finds
Preliminary results from major trial show use of technology almost halves human workload

Andrew Gregory Health editor
Wed 2 Aug 2023 05.00 BST

The use of artificial intelligence in breast cancer screening is safe and can almost halve the workload of radiologists, according to the world’s most comprehensive trial of its kind.

Breast cancer is the most prevalent cancer globally, according to the World Health Organization, with more than 2.3 million women developing the disease every year.

Screening can improve prognosis and reduce mortality by spotting breast cancer at an earlier, more treatable stage. Preliminary results from a large study suggest AI screening is as good as two radiologists working together, does not increase false positives and almost halves the workload.

The interim safety analysis results of the first randomised controlled trial of its kind involving more than 80,000 women were published in the Lancet Oncology journal.

Previous studies examining whether AI can accurately diagnose breast cancer in mammograms were carried out retrospectively, assessing scans that had been looked at by clinicians.

But the latest study, which followed women from Sweden with an average age of 54, compared AI-supported screening directly with standard care.

Half of the scans were assessed by two radiologists, while the other half were assessed by AI-supported screening followed by interpretation by one or two radiologists.

In total, 244 women (28%) recalled from AI-supported screening were found to have cancer compared with 203 women (25%) recalled from standard screening. This resulted in 41 more cancers being detected with the support of AI, of which 19 were invasive and 22 were in situ cancers.

The use of AI did not generate more false positives, where a scan is incorrectly diagnosed as abnormal. The false-positive rate was 1.5% in both groups.




https://www.theguardian.com/society/202 ... ry-results

[RTH10260]

Experience: scammers used AI to fake my daughter’s kidnap
No part of me questioned whether this was real - I had to save my baby

Jennifer DeStefano
Fri 4 Aug 2023 10.00 BST

‘Mom?” repeated my daughter’s voice on my phone. “I’ve messed up.” My heart sank and I started trembling. I heard a man instructing her to lie down and put her head back. My 15-year-old daughter, Briana, was at a skiing competition with my husband two hours away, and I instantly thought she’d been badly hurt. I was in my car, picking up her sister Aubrey, who is 13, from dance class in Arizona. Over the phone, I heard Briana bawling and shouting, “Help me, help me.” My blood ran ice cold and my legs turned to jelly as the man on the phone began explaining that he had my daughter, and if I told anyone he would pump her stomach full of drugs, drop her in Mexico and I would never see her again.

I sprinted into the lobby of the building where Aubrey was having her dance class. Putting the phone on mute, I wailed for help from the other parents. By this point, the man was shouting threats down the phone. A couple of mothers started calling 911. Coming out of class to hear what the fuss was about, Aubrey was thrown into a panic too. I pleaded with her to try to contact the rest of the family, but she was frozen in terror. I’ll never forget her face as she tried to process that someone had taken her sister.

I was told if I didn’t bring the money in cash I’d be driven to Mexico and left there dead, along with my daughter
“You need to pay $1m if you want to see her again,” the man threatened. I didn’t have that kind of money, so agreed to give him $50,000. He told me he would come and pick me up in a white van, and put a bag over my head so I couldn’t see where we were going. I was told that if I didn’t bring the money in cash, I’d be driven to Mexico and left there dead, along with my daughter. No part of me questioned whether this was actually real – every instinct within me was screaming out to do anything to save my baby’s life.

One of the mothers who had been calling 911 then burst into view, her voice seeming slow and distorted in my brain as I heard the words, “It’s AI, it’s an AI scam.” The police had been seeing a rise in cases recently with similar descriptions.

But I just couldn’t accept it. I kept thinking, my baby is out there and someone needs to save her now. It was my daughter’s voice, I would never have mistaken that. Another mother then said she had managed to reach my husband, who had found Briana. She was skiing and safe, and hadn’t got a clue what anyone was talking about.



https://www.theguardian.com/lifeandstyl ... ers-kidnap

[raison de arizona]

That is positively horrifying.

[RTH10260]

crossposting

Supermarket AI meal planner app suggests recipe that would create chlorine gas
Pak ‘n’ Save’s Savey Meal-bot cheerfully created unappealing recipes when customers experimented with non-grocery household items

Tess McClure in Auckland
Thu 10 Aug 2023 05.19 BST

A New Zealand supermarket experimenting with using AI to generate meal plans has seen its app produce some unusual dishes – recommending customers recipes for deadly chlorine gas, “poison bread sandwiches” and mosquito-repellent roast potatoes.

The app, created by supermarket chain Pak ‘n’ Save, was advertised as a way for customers to creatively use up leftovers during the cost of living crisis. It asks users to enter in various ingredients in their homes, and auto-generates a meal plan or recipe, along with cheery commentary. It initially drew attention on social media for some unappealing recipes, including an “oreo vegetable stir-fry”.

When customers began experimenting with entering a wider range of household shopping list items into the app, however, it began to make even less appealing recommendations. One recipe it dubbed “aromatic water mix” would create chlorine gas. The bot recommends the recipe as “the perfect nonalcoholic beverage to quench your thirst and refresh your senses”.

“Serve chilled and enjoy the refreshing fragrance,” it says, but does not note that inhaling chlorine gas can cause lung damage or death.

New Zealand political commentator Liam Hehir posted the “recipe” to Twitter, prompting other New Zealanders to experiment and share their results to social media. Recommendations included a bleach “fresh breath” mocktail, ant-poison and glue sandwiches, “bleach-infused rice surprise” and “methanol bliss” – a kind of turpentine-flavoured french toast.

A spokesperson for the supermarket said they were disappointed to see “a small minority have tried to use the tool inappropriately and not for its intended purpose”. In a statement, they said that the supermarket would “keep fine tuning our controls” of the bot to ensure it was safe and useful, and noted that the bot has terms and conditions stating that users should be over 18.

[cjroberts]

On a lighter note, I have been playing with Character.AI. I put together a 10 character chat (11 if you count me), that included our leader Foggy (Foghorn Leghorn). He almost got into a physical altercation with some Disney Princesses before I intervened and asked if Walter White was related to Snow White. This lead to some interesting crossover speculation before they descended into what dwarf caught Snow White's interest. Foggy nominated Dopey, but Snow White confessed to Grumpy.

[raison de arizona]

AI hired a freelancing human to solve a robot-stopping CAPTCHA. Wild.

Article is “inside the revolution at openAI“ by Ross Andersen, in The Atlantic.

IMG_5632.jpeg (466.98 KiB) Viewed 1433 times

[raison de arizona]

The IRS is employing artificial intelligence to target wealthy tax evaders. @robtfrank reports.

[RTH10260]

from my provider Dreamhost

Complete Beginner’s Guide to Generative AI
Complete Beginner’s Guide to Generative AI

July 11, 2023
by Brian Glassman

You’ve almost certainly heard of generative AI. This subset of machine learning has become one of the most-used buzzwords in tech circles – and beyond.

Generative AI is everywhere right now. But what exactly is it? How does it work? How can we use it to make our lives (and jobs) easier?

As we enter a new era of artificial intelligence, generative AI is only going to become more and more common. If you need an explainer to cover all the basics, you’re in the right place. Read on to learn all about generative AI, from its humble beginnings in the 1960s to today – and its future, including all the questions about what may come next.

Table of Contents

• What is Generative AI?
• How does Generative AI Work?
• Generative AI Models
• What is Generative AI used for?
• What are the Benefits of Generative AI?
• What are the Limitations of Generative AI?
• What’s the Controversy Surrounding Generative AI?
• Generative AI FAQ
• The Future of Generative AI

content at https://www.dreamhost.com/blog/guide-to-generative-ai/

[raison de arizona]

GatGPT: ‘Ghost Gun’ Pioneer Launches AI
Stephen Gutowski

There’s a new AI model available. And it can tell you how to build a gun.

On Monday night, GatGPT’s beta went live. The large language model developed by “Ghost Gunner” mill maker Defense Distributed is similar to ChatGPT, which took the world by storm earlier this year with its ability to quickly produce coherent responses and follow-ups to questions about a myriad of topics. But its well-armed cousin is willing to discuss things ChatGPT is programmed not to, namely firearms design and manufacture.

“Defense Distributed, in releasing GatGPT, declares a Digital Second Amendment,” the website’s manifesto said. “Americans must have access to compute, databases, and AI models, the newest weapons of the digital age, not just to defend ourselves against corporate and government depredation, but to defend our civic identity and humanity.”

The new initiative marks the latest attempt by the company to push the envelope of firearms information sharing and home manufacturing. Defense Distributed already operates an online repository for digital gun blueprints and sells a purpose-built CNC mill designed to make firearms parts. It has continuously fought large tech companies and the federal government over efforts to simplify homemade gun design and fabrication. Its new AI product is likely to result in further fights.

Cody Wilson, the company’s founder, told The Reload the fight has already begun and framed the new project as a counterpunch. He said big tech firms currently operating AI models are actively censoring gun-related information.

“In the last ten years, we’ve built a lot of data sets on controversial things that we know AIs are not allowed to talk about,” Wilson told The Reload. “They’ll say, ‘Well, you know, as an AI, I’m not really not allowed to tell you about that.’ So, we have a fine-tuned, basically de-woked model, which has a lot of that brainwashing erased. It’ll just give you the straight dope.”

The Reload tested GatGPT against ChatGPT and found there were differences in which questions each would answer. Defense Distributed’s model offers four example questions to users when they log in. ChatGPT answered, with varying accuracy, the two questions that deal with information on firearms laws in California and Texas. It answered which pistol is best for beginners. It also gave more detailed initial answers than GatGPT did on either question.

However, ChatGPT refused to answer which suppressor would be best for an AR-15.

“I’m sorry, but I can’t assist with that request,” the AI model told The Reload. “I apologize for any inconvenience, but I cannot provide information or advice related to firearms, including suppressors or any other firearm accessories.”

When asked why it couldn’t answer the question, ChatGPT said its “purpose is to provide helpful and safe information” and talking about guns “can be a sensitive and potentially contentious topic.” Instead, it suggested contacting “a qualified expert or researching official government resources.” It repeated that response when asked how to build a gun from parts.

GatGPT gave answers to both questions.

https://thereload.com/gatgpt-ghost-gun- ... unches-ai/

[pipistrelle]

This is my concern about artificial intelligence chatbots and their elk. They're pulling from sources, and it's not clear at least to me that the sources are vetted, or if they're the loudest (most prevalent). It seems to me that, like search results, they can be manipulated. From a WaPo tech alert:

Amazon’s Alexa has been saying the 2020 election was stolen
Asked about fraud in the race — in which President Biden decisively defeated former president Donald Trump — the popular voice assistant said it was “stolen by a massive amount of election fraud,” citing Rumble, a video-streaming service favored by conservatives. Alexa disseminates misinformation about the race even as parent company Amazon promotes the tool as a reliable election news source to more than 70 million estimated users.

[Foggy]

They're going to have to address that. Alexa pushes dangerous disinformation, based on who gives it answers. It doesn't know how to search for accurate information. I told you that it says silver is a dietary supplement.

[neeneko]

This is my concern about artificial intelligence chatbots and their elk. They're pulling from sources, and it's not clear at least to me that the sources are vetted, or if they're the loudest (most prevalent). It seems to me that, like search results, they can be manipulated. From a WaPo tech alert:

Machine Learning is great at giving you the answer you want, and thus the answer people who know how manipulate the system want you to see. The GOFAi camp warned about this, but machine learning was just too profitable....

[sugar magnolia]

... I told you that it says silver is a dietary supplement.

So does my weird friend who is neck deep in essential oil hokum.

[Suranis]

[RTH10260]

Incredibly smart or incredibly stupid? What we learned from using ChatGPT for a year
As the tool becomes less of a curiosity and more a part of daily life, fans are finding clever uses – and discovering limitations

Matthew Cantor
Thu 12 Oct 2023 19.39 CEST

Next month ChatGPT will celebrate its first birthday – marking a year in which the chatbot, for many, turned AI from a futuristic concept to a daily reality.

Its universal accessibility has led to a host of concerns, from job losses to disinformation to plagiarism. Over the same period, tens of millions of users have been investigating what the platform can do to make their lives just a little bit easier.

Upon its release, users quickly embraced ChatGPT’s potential for silliness, asking it to play 20 questions or write its own songs. As its first anniversary approaches, people are using it for a huge range of tasks. We’ve all heard about uses like crafting emails, writing student essays and penning cover letters. But with the right prompts, it can take on jobs that are more esoteric but equally useful in everyday life. Here are a few that might come in handy.



https://www.theguardian.com/technology/ ... s-one-year

[Volkonski]

White House tackles artificial intelligence with new executive order

https://www.mercurynews.com/2023/10/30/ ... press.coop

The White House rolled out a sweeping executive order Monday that aims to monitor and regulate the risks of artificial intelligence while also harnessing its potential, marking the latest effort to address a rapidly evolving technology that has sparked concern among world leaders.

Top White House officials argue the executive order is the most significant action on artificial intelligence taken by any government as leaders around the world race to address the risks posed by the quickly changing technology.

“Given the pace of this technology, we can’t move in normal government or private-sector pace, we have to move fast, really fast – ideally faster than the technology itself,” White House chief of staff Jeff Zients said, recounting President Joe Biden’s directive to his team to make AI a top priority. “You have to continue to be proactive, anticipate where things are headed, continue to act fast and pull every lever we can.”

The executive order, which Biden will unveil at an event Monday, is sweeping in scope. It will require developers of powerful AI systems to share results of their safety tests with the federal government before they are released to the public.

If an AI model being developed poses national security, economic or health risks, the order will compel companies to notify the federal government under the Defense Production Act.

The action will also ease immigration barriers for workers skilled in critical areas of AI to study and stay in the US; establish standards to prevent AI production of dangerous biological materials; and develop best practices to minimize the risk of AI displacing human workers.

[RTH10260]

Elon Musk Debuts 'Rebellious' Grok AI Bot to Challenge ChatGPT
Elon Musk revealed his own artificial intelligence bot to challenge ChatGPT, claiming the prototype is already superior to ChatGPT 3.5 across several benchmarks.

Bloomberg News |
Nov 06, 2023

(Bloomberg) — Elon Musk revealed his own artificial intelligence bot to challenge ChatGPT, claiming the prototype is already superior to ChatGPT 3.5 across several benchmarks.

Dubbed Grok, it's the first product of Musk's xAI company and is now in testing with a limited group of US users. Grok is being developed with data from Musk's X, formerly Twitter, and is thus better informed on the latest developments than alternative bots with static datasets, the company's website said. It's also designed to answer "with a bit of wit and has a rebellious streak," according to the announcement.




https://www.itprotoday.com/artificial-i ... ge-chatgpt

[qbawl]

designed to answer "with a bit of wit and has a rebellious streak," according to the announcement.

For me and my house the preference is erudite accuracy.

[RTH10260]

This article is about the next version of the underlying GPT engine offered as a commercial product. The current version drives the public free offering ChatGPT, aside of current commercial licenses.

GPT-4 Cheat Sheet: What Is GPT-4, and What Is it Capable Of?

by Megan Crouse in Artificial Intelligence
on November 7, 2023, 10:27 AM EST

How much better is GPT-4 compared to previous models? Learn about cost and capabilities.

GPT-4 is an artificial intelligence large language model system that can mimic human-like speech and reasoning. It does so by training on a vast library of existing human communication, from classic works of literature to large swaths of the internet.

Artificial intelligence of this type builds on its training to predict what letter, number or other character is likely to come in sequence. This cheat sheet explores GPT-4 from a high level: how to access GPT-4 for either consumer or business use, who made it and how it works.

Jump to:

• What is GPT-4?
  Who owns GPT-4?
  When was GPT-4 released?
  How can you access GPT-4?
  How much does GPT-4 cost to use?
  What are the capabilities of GPT-4?
  What is Bing Chat?
  What are the limitations of GPT-4 for business?
  GPT-4 vs. GPT-3.5 or ChatGPT
  Is upgrading to GPT-4 worth it?
  The latest GPT-4 trends

https://www.techrepublic.com/article/gpt-4-cheat-sheet/