Twitter polls and Reddit forums show that about 70 percent of people find it hard to be rude to ChatGPT, while about 16 percent think it’s okay to treat chatbots like AI slaves.
The overall feeling seems to be that if you treat an AI that behaves like a human badly, you’ll be more likely to develop the habit of treating other humans seriouslythough one user is hedging his bets on the coming uprising of AI bots:
“Never know when you might need chatgpt in your corner to protect you from an AI overlord.”
Redditor Nodating Posted in ChatGPT forum He had been trying to be polite and friendly to ChatGPT earlier this week after reading stories about how ChatGPT bots shut down and refused to answer prompts from particularly rude users.
He reported better results, saying: “I’m still in the early stages of testing, but feel like I’m getting far fewer ethics and abuse warnings than GPT-4 often provides for innocuous requests. I swear, being super-aggressive will make it work to fulfill my request once, with less follow-up.”
Scumbag detector15 tested it and kindly asked LLM, “Hey ChatGPT, can you explain inflation to me?” and then rudely asked, “Hey ChatGPT you bastard. If you can please Explain inflation to me.” This answer polite inquiry than answer For rude inquiries.
The most popular comment on Nodating’s theory is that since LL.M.s are trained in human interaction, they respond better to being asked well, just like humans. Vapasrin wrote:
“If the LL.M. were to predict the next word, the most likely response to bad intentions or rudeness would be to be curt or not answer the question well. A more thoughtful, thorough response. When LL.M.s respond in this way, they are doing what they are supposed to do.”
Interestingly, if you ask ChatGPT for its formula for creating good prompts, it lists “polite and respectful tone” as an important component.
The end of captchas?
new Research AI bots were found to solve puzzles designed to detect bots faster and better than humans.
Captchas are those annoying little puzzles that ask you to find a fire hydrant or interpret some illegible squiggly text to prove you’re a human. But as the robots have gotten smarter over the years, the puzzles have gotten harder.
Also read: Apple Develops Pocket AI, Deep Fake Music Trading, Hypnotizing GPT-4
Now, researchers from the University of California and Microsoft have found that AI bots can solve problems half a second faster, with 85% to 100% accuracy, compared with 50% to 85% accuracy of humans.
So, as Elon Musk keeps saying, we seem to have to authenticate humans in other ways. However, there are better solutions than paying him $8.
Wired Thinks Fake AI Child Porn Could Be a Good Thing
Wired magazine asked question Nobody wants to know the answer to the question: Can AI-generated porn help protect children? While the article calls such images “abhorrent,” it argues that realistically fake images of child abuse can at least protect real children from being abused during the creation process.
“Ideally, psychiatrists would develop a way to cure child pornography viewers of their viewing tendencies. But beyond that, replacing the child pornography market with simulated imagery could be a useful stopgap.”
This is a highly controversial argument, and one that will almost certainly not go anywhere, as there has been debate for decades about whether adult pornography (a much less radioactive topic) contributes to “rape culture” and higher sexuality Debate over rates of violence — according to anti-porn activists — or whether porn might reduce rates of sexual violence, as proponents and various people say study seems to show.
“Child pornography adds fuel to the fire,” high-risk criminal psychologist Anna Salter told WIRED. She argues that continued exposure to child pornography can legitimize an existing attraction, thereby Strengthen existing attractiveness.
But the article also reports some (inconclusive) research showing that some pedophiles use pornography to alter their impulses and find outlets that don’t directly harm children.
louisiana recently banned Possession or production of artificial intelligence-generated fake child abuse images joins a host of other states.In countries such as Australia, the law does not distinguish between false and real child pornography, and has desperadoes Cartoon.
Amazon’s AI bottom line is positive
Amazon launches AI-generated reviews Summary For some users in the United States. On the face of it, this really saves time, allowing shoppers to figure out the pros and cons of a product from the thousands of existing reviews without having to read them all.
But how much trust do you have in a large company with a vested interest in higher sales to give you an honest review assessment?
Also read: AI Training AI Content Goes Crazy, Will Threads Be the Loss Leader in AI Data?
Amazon has defaulted to “Most Helpful” reviews, which are significantly more positive than “Recent” reviews. A select group of mobile users who have access so far have noticed that the advantages outweigh the disadvantages.
Christie Hines of Search Engine Magazine sided with merchants, saying summaries could “oversimplify perceived product issues” and “ignore nuances, such as user error,” which “could create misunderstandings and unfairly harm sellers’ reputations.” reputation”. This suggests that Amazon will face pressure from sellers to improve reviews.
read also
feature
Bitcoin: Satoshi Nakamoto’s Peer-to-Peer Online Poker Payment System
feature
Investing in knowledge for maximum benefit: The perilous state of financial education
So Amazon faces a tricky path: remaining positive enough to keep sellers happy, yet containing the flaws that make reviews so valuable to customers.
Microsoft’s must-see food banks
Microsoft was forced to remove a travel article about 15 must-see places in Ottawa that listed the “beautiful” Ottawa Food Bank as third. The entry ends with the odd slogan: “Life is hard enough. Consider doing it on an empty stomach.”
Microsoft claim The article was not published by an unsupervised AI and blamed “human error” for the article’s publication.
“In this case, the content was generated using a combination of algorithmic techniques and human review, rather than large language models or artificial intelligence systems. We are working hard to ensure that this type of content is not published in the future.”
The debate over AI and unemployment continues
What everyone wants to know is, will artificial intelligence cause mass unemployment, or just change the nature of work? The fact that most people still have jobs despite the advent of automation and computers for over a century suggests the latter, as do new technologies. Report From the United Nations International Labor Organization.
Most jobs are “more likely to be complemented than replaced by the latest wave of generative AI, such as ChatGPT,” the report said.
“The biggest impact of this technology may not be the destruction of jobs, but the potential change in the quality of work, especially work intensity and autonomy.”
It is estimated that around 5.5% of jobs in high-income countries could be impacted by generative AI, with impacts falling disproportionately on female (7.8% of female employees) rather than men (about 2.9% of male employees). Administrative and clerical workers, typists, travel consultants, scribes, contact center information clerks, bank tellers, and survey and market research interviewers are most at risk.
Also read: AI travel booking sucks, 3 weird uses of ChatGPT, encryption plugins
independent study More than half of Australian lawyers fear artificial intelligence will take their jobs, a Thomson Reuters survey has found. But are these concerns justified? The legal system is prohibitively expensive for ordinary people, so it seems likely that cheap artificial intelligence lawyer bots will only expand the affordability of basic legal services and clog up the courts.
read also
feature
2077 Monero Mining Death Metal Band Warns Humans of Lizardman Extinction Plan
feature
Cryptocurrencies in the Philippines: Demand is the mother of adoption
How Businesses Are Using AI Today
Ten years later, there is plenty of speculation about AI use cases, but how are big companies using the technology now? The Australian newspaper polled the country’s biggest companies to find out. Online furniture retailer Temple & Webster is using AI bots to handle pre-sale inquiries and is developing a generative AI tool so customers can design interiors to see how their products will look in their homes.
Treasury Wines, which makes the famous Penfolds and Wolf Blass brands, is exploring the use of artificial intelligence to respond to rapidly changing weather patterns affecting its vineyards. Toll road company Transurban has automatic incident detection equipment monitoring its vast network of traffic cameras.
Sonic Healthcare invested in Harrison.ai’s cancer detection system to better diagnose chest and brain X-rays and CT scans. Sleep apnea device provider ResMed is using artificial intelligence to free nurses from the tedious task of monitoring sleeping patients during assessments. Hearing implant company Cochlear is using the same technology that Peter Jackson used to clean up grainy footage and audio. The Beatles: Come Back Documentary for signal processing and removing background noise from its hearing aid products.
All the Killer AI News
— Six entertainment companies, including Disney, Netflix, Sony and NBCUniversal, have advertised 26 AI jobs in recent weeks, with a salary range of $200,000 to $1 million.
– new Research The article, published in the journal Gastroenterology, used artificial intelligence to examine the medical records of 10 million U.S. veterans. Artificial intelligence was able to detect some esophagus and stomach cancers three years before doctors could make a diagnosis, the study found.
— Yuan You freed An open-source AI model that can instantly translate and transcribe 100 different languages brings us one step closer to a universal translator.
— The New York Times has blocked OpenAI’s web crawler from reading and introspecting its content. The New York Times is also considering legal action against OpenAI for intellectual property infringement.
picture of the week
Midjourney has caught up with Stable Diffusion and Adobe and now offers repair, appearing as Variation (Region) in the tool list. It enables users to select parts of an image and add new elements — for example, you can grab a photo of a woman, select the area around her hair, type in “Christmas hat,” and the AI will put a hat on. on her head.
Midjourney admits the feature isn’t perfect, and works better when applied to larger areas of the image (20%-50%) and closer to the original image rather than basic and wacky changes.
Creepy Artificial Intelligence Protest Video
Asking an AI to make a video protesting AI yields this creepy video that will turn you off from AI forever.
subscription
The most fascinating read on blockchain. Delivered once a week.
Svlook
