## Turing test passed again and again treated as a prank.

The war between wetware and hardware.
Giz
Posts: 5188
Joined: Mon Jul 12, 2004 5:07 pm
Location: USA!

### Re: Turing test passed again and again treated as a prank.

Anaxagoras wrote: Thu Oct 11, 2018 9:18 am
Rob Lister wrote: Thu Oct 11, 2018 9:09 am
Anaxagoras wrote: Wed Oct 10, 2018 11:55 pm That's funny but I think it was written by a human. Or at least the jokes in it were. Somehow it was programmed to write something similar. A flag that says "Arby's food is fine to eat" at a Trump rally? Why would a computer think that Arby's food is funny? "Great jobs. Tall jobs. Steve Jobs." That one sort of does follow a humor formula, the "one of these things is not like the others" formula.
https://me.me/i/keaton-patti-keatonpatt ... 6a6576fcd5

I don't think it was meant as humor. It just is -ous.
Yes. I just have my doubts that it was written by a bot.
Turing test passed again but treated as a prank?
ed
Posts: 42122
Joined: Tue Jun 08, 2004 11:52 pm
Title: G_D

### Re: Turing test passed again and again treated as a prank.

Could Rob be a Bot? It would explain much. Pillory might be V.0.8
xouper
Posts: 11741
Joined: Fri Jun 11, 2004 4:52 am
Title: mere ghost of his former self

### Re: Turing test passed again and again treated as a prank.

Abdul Alhazred wrote: Wed Oct 17, 2018 2:04 pm
ed wrote: Wed Oct 17, 2018 1:55 pm Could Rob be a Bot? It would explain much. Pillory might be V.0.8
Everybody here is a 'bot, except Pyrrho, me, and maybe Doctor X. :P
Tell me more about Pyrrho and you and maybe Doctor X.
Witness
Posts: 35689
Joined: Thu Sep 19, 2013 5:50 pm

### Re: Turing test passed again and again treated as a prank.

AI tags potential criminals before they’ve done anything

British police want to use AI for highlighting who is at risk of becoming a criminal before they’ve actually committed any crime.

Although it sounds like a dystopian nightmare, there are clear benefits.

Resources and outreach programs can be allocated to attempt preventing a crime, stop anyone becoming a victim, and remove the costs associated with prosecuting and jailing someone.
[…]
The proposed system is called the National Data Analytics Solution (NDAS) and uses data from local and national police databases.

According to NDAS’ project leader, over a terabyte of data has been collected from the aforementioned databases. This data includes logs of committed crimes in addition to around five million identifiable people.

There are over 1,400 indicators within the data which the AI uses to calculate an individual’s risk of committing a crime. Such indicators could include past offences, whether the person had assistance, and whether those in their network are criminals.
https://www.artificialintelligence-news ... criminals/

:notsure:
Giz
Posts: 5188
Joined: Mon Jul 12, 2004 5:07 pm
Location: USA!

### Re: Turing test passed again and again treated as a prank.

Was Minority Report a cautionary tale, or a blueprint?

Presumably, being the UK, this isn’t actually anything useful. It’s just predicting who’s likely to write mean tweets.
Anaxagoras
Posts: 30086
Joined: Wed Mar 19, 2008 5:45 am
Location: Yokohama/Tokyo, Japan

### Re: Turing test passed again and again treated as a prank.

ten quid says they will eventually be "shocked" to discover that this AI is "racist". ;)
gnome
Posts: 25992
Joined: Tue Jun 29, 2004 12:40 am
Location: New Port Richey, FL

### Re: Turing test passed again and again treated as a prank.

Abdul Alhazred wrote: Wed Jan 02, 2019 5:40 pm This clever AI hid data from its creators to cheat at its appointed task
Tech Crunch
Depending on how paranoid you are, this research from Stanford and Google i will be either terrifying or fascinating. A machine learning agent intended to transform aerial images into street maps and back was found to be cheating by hiding information it would need later in “a nearly imperceptible, high-frequency signal.”

...

In some early results, the agent was doing well — suspiciously well. What tipped the team off was that, when the agent reconstructed aerial photographs from its street maps, there were lots of details that didn’t seem to be on the latter at all. For instance, skylights on a roof that were eliminated in the process of creating the street map would magically reappear when they asked the agent to do the reverse process ...
Horseshit.

My money's on that a human being specifically and possibly secretly instructed the agent to do this, in order to make it appear to perform better than it otherwise could. If it was literally the "machine learning" part of the agent that led it to devise that method, I'll ... something. Someone name a dumb avatar for me to pick for 17 days.
Witness
Posts: 35689
Joined: Thu Sep 19, 2013 5:50 pm

### Re: Turing test passed again and again treated as a prank.

Microsoft Invests in ‘Boring AI’ Startup Valued at $2.75 Billion Databricks Inc., which says it makes software to help companies satisfy their “boring AI” needs, raised a new round of funding that values the business at$2.75 billion.

Microsoft Corp. was among the investors in the $250 million financing, the company said on Tuesday. […] Databricks sells tools that pool a company’s databases and apply a form of artificial intelligence to search for meaningful information. […] Ali Ghodsi, chief executive officer of Databricks, said he plans to use the new capital to hire, open new offices and make headway into other industries, such as media, retail and government. Part of his pitch: Databricks “satisfies the needs of big enterprise customers that want to do boring AI use cases.” https://www.bloomberg.com/news/articles ... 75-billion The robotic overlords disguise so well they appear boring. :o Witness Posts: 35689 Joined: Thu Sep 19, 2013 5:50 pm ### Re: Turing test passed again and again treated as a prank. New AI fake text generator may be too dangerous to release, say creators The Elon Musk-backed nonprofit company OpenAI declines to release research publicly for fear of misuse The creators of a revolutionary AI system that can write news stories and works of fiction – dubbed “deepfakes for text” – have taken the unusual step of not releasing their research publicly, for fear of potential misuse. OpenAI, an nonprofit research company backed by Elon Musk, Reid Hoffman, Sam Altman, and others, says its new AI model, called GPT2 is so good and the risk of malicious use so high that it is breaking from its normal practice of releasing the full research to the public in order to allow more time to discuss the ramifications of the technological breakthrough. https://www.theguardian.com/technology/ ... ws-fiction Anaxagoras Posts: 30086 Joined: Wed Mar 19, 2008 5:45 am Location: Yokohama/Tokyo, Japan ### Re: Turing test passed again and again treated as a prank. Bullshit. but clever bullshit. Pyrrho Posts: 33281 Joined: Sat Jun 05, 2004 2:17 am Title: Man in Black Location: Division 6 ### Re: Turing test passed again and again treated as a prank. Link: xouper Posts: 11741 Joined: Fri Jun 11, 2004 4:52 am Title: mere ghost of his former self ### Re: Turing test passed again and again treated as a prank. Pyrrho wrote: Thu Mar 21, 2019 10:53 am Link: Thanks. That made me chuckle. That's a clever joke. It's clever because it is totally accurate and yet totally misses the point, which is presumably intentional. You could just as easily go one level deeper and say: browser extension idea: find and replace “AI” with “a sequence of ones and zeros“ in articles you read online ie. VCs invest$1.2B in a sequence of ones and zeros

Similarly:
browser extension idea: find and replace “airplane” with “an assemblage of metal and plastic parts“ in articles you read online

ie. airline company invests \$1.2B in an assemblage of metal and plastic parts

The cleverness of the original joke is in the choice of the substitute phrase. Choose a good one, as he did, and it's funny. Choose a bad one, as I did, and it's lame.
Witness
Posts: 35689
Joined: Thu Sep 19, 2013 5:50 pm

### Re: Turing test passed again and again treated as a prank.

↑ The article doesn't say that a false positive can also get you killed. :|
gnome
Posts: 25992
Joined: Tue Jun 29, 2004 12:40 am
Location: New Port Richey, FL

### Re: Turing test passed again and again treated as a prank.

When has being mistaken for a criminal ever gotten someone killed? Except all those times it happened.
Anaxagoras
Posts: 30086
Joined: Wed Mar 19, 2008 5:45 am
Location: Yokohama/Tokyo, Japan

### Re: Turing test passed again and again treated as a prank.

That's OK, I'm sure eventually they'll work out the kinks. Hey, nobody's perfect, right? :wink:

Side anecdote: I was watching a TV show on netflix the other day, and in one scene, this character appears in a wig that is a different color and hairstyle from her normal hair. She also had different makeup and clothes. Although the scene doesn't explicitly tell you that this is the main character, it was strongly implied from the context, but I had to pause it and look carefully at her face to recognize her. In real life, if I didn't have some other clue, the disguise would have totally fooled me. I guess I'm just not very good at "facial recognition" myself.
Rob Lister
Posts: 23535
Joined: Sun Jul 18, 2004 7:15 pm
Title: Incipient toppler
Location: Swimming in Lake Ed

### Re: Turing test passed again and again treated as a prank.

Wrongly? Are you sure?
Pyrrho
Posts: 33281
Joined: Sat Jun 05, 2004 2:17 am
Title: Man in Black
Location: Division 6

### Re: Turing test passed again and again treated as a prank.

robinson
Posts: 19621
Joined: Sat Aug 12, 2006 2:01 am
Title: Je suis devenu Français
Location: USA

### Re: Turing test passed again and again treated as a prank.

Reddit is what happens when almost anybody or anything can post. Almost anything.

4chan is what happens when you can post anything
Anaxagoras
Posts: 30086
Joined: Wed Mar 19, 2008 5:45 am
Location: Yokohama/Tokyo, Japan

### Re: Turing test passed again and again treated as a prank.

Abdul Alhazred wrote: Sat Oct 10, 2020 1:32 pm Someone not only created a comment-spewing Reddit bot powered by OpenAI's GPT-3, it offered bizarre life advice
The Register
Someone used OpenAI's GPT-3 text-generating software to write a string of posts on Reddit, convincing people the missives were penned by a real person, and banking internet points in the process, The Register can confirm.

Redditor thegentlemetre has produced over the course of the past 10 days hundreds of answers on the popular subreddit AskReddit in which people turn to millions of their peers for practical tips and life advice. At its most frenzied periods, the account generated about one comment per minute, offering a jumble of answers to people's questions. That would be quite a feat for a lone human, considering most of these submissions are several paragraphs long.

Some of its responses are incoherent and downright creepy, such as this fake story about a colony of humans living trapped underneath elevators. Sometimes it doesn't answer a netizen's question at all, and sometimes it spreads false information, such as denying that the state of Nebraska exists. There's one post where it believes men are discriminated against when it comes to being topless in public.

...
Ether it passed the Turing test, or thousands of Redditors fail it every day. :)
Most of the other commenters apparently sussed out that it was a bot. Unfortunately, the mean moderators removed its comment, which is too bad because I wanted to read about the colony of humans living trapped underneath elevators.

Anaxagoras
Posts: 30086
Joined: Wed Mar 19, 2008 5:45 am
Location: Yokohama/Tokyo, Japan

### Re: Turing test passed again and again treated as a prank.

A.I. being sexist!!! :x :o

When AI Sees a Man, It Thinks 'Official.' A Woman? 'Smile'
A new paper renews concerns about bias in image recognition services offered by Google, Microsoft, and Amazon.
Some interesting things in the article, besides the obvious.
The researchers administered their machine vision test to Google’s artificial intelligence image service and those of rivals Amazon and Microsoft. Crowdworkers were paid to review the annotations those services applied to official photos of lawmakers and images those lawmakers tweeted.
Google switched off its AI vision service’s gender detection earlier this year, saying that gender cannot be inferred from a person’s appearance. Tracy Frey, managing director of responsible AI at Google’s cloud division, says the company continues to work on reducing bias and welcomes outside input. “We always strive to be better and continue to collaborate with outside stakeholders—like academic researchers—to further our work in this space,” she says. Amazon and Microsoft declined to comment; both companies’ services recognize gender only as binary.
Didn't make any difference apparently. If anything, Google's AI seemed to be more "biased" than the other two (although they all exhibited some biases).
All 20 lawmakers are smiling in their official photos. Google’s top suggested labels noted a smile for only one of the men, but for seven of the women. The company’s AI vision service labeled all 10 of the men as “businessperson,” often also with “official” or “white collar worker.” Only five of the women senators received one or more of those terms. Women also received appearance-related tags, such as “skin,” “hairstyle,” and “neck,” that were not applied to men.

Amazon and Microsoft’s services appeared to show less obvious bias, although Amazon reported being more than 99 percent sure that two of the 10 women senators were either a “girl” or “kid.” It didn’t suggest any of the 10 men were minors. Microsoft’s service identified the gender of all the men, but only eight of the women, calling one a man and not tagging a gender for another.
robinson
Posts: 19621
Joined: Sat Aug 12, 2006 2:01 am
Title: Je suis devenu Français
Location: USA

### Re: Turing test passed again and again treated as a prank.

Why would anyone give a shit?
xouper
Posts: 11741
Joined: Fri Jun 11, 2004 4:52 am
Title: mere ghost of his former self

### Re: Turing test passed again and again treated as a prank.

I didn't know which of the AI threads to put this in, so I asked Alexa. Thus, if this is the wrong thread, it's Rohit Prasad's fault.

• The Turing Test is obsolete. It’s time to build a new barometer for AI
By Rohit Prasad, December 28, 2020
https://www.fastcompany.com/90590042/tu ... azon-alexa
This year marks 70 years since Alan Turing published his paper introducing the concept of the Turing Test in response to the question, “Can machines think?”

. . . Turing predicted that by the year 2000, an average human would have less than a 70% chance of distinguishing an AI from a human . . .

. . . Why haven’t we as an industry been able to achieve that goal, 20 years past that mark? I believe the goal put forth by Turing is not a useful one for AI scientists like myself to work toward.

. . . None of this is to denigrate Turing’s original vision — Turing’s “imitation game” was designed as a thought experiment, not as the ultimate test for useful AI. However, now is the time to dispel the Turing Test and get inspired by Alan Turing’s bold vision to accelerate progress in building AIs that are designed to help humans.

Rohit Prasad is vice president and head scientist of Alexa at Amazon.

A commentary on the above commentary:
• Is the Turing Test Obsolete?
By Joel Hruska, January 4, 2021
https://www.extremetech.com/computing/3 ... t-obsolete
Prassad is absolutely right that the Turing Test has acknowledged limitations. It tests whether a computer behaves like a human being, not whether a computer demonstrates something we might call “intelligence.”

A commentary on the commentary of the commentary:
As always, Your Mileage May Vary™.
Anaxagoras
Posts: 30086
Joined: Wed Mar 19, 2008 5:45 am
Location: Yokohama/Tokyo, Japan

### Re: Turing test passed again and again treated as a prank.

xouper wrote: Tue Jan 05, 2021 1:58 am I didn't know which of the AI threads to put this in, so I asked Alexa. Thus, if this is the wrong thread, it's Rohit Prasad's fault.

• The Turing Test is obsolete. It’s time to build a new barometer for AI
By Rohit Prasad, December 28, 2020
https://www.fastcompany.com/90590042/tu ... azon-alexa
With AI now ubiquitously integrated into our phones, cars, and homes, it’s become increasingly obvious that people care much more that their interactions with machines be useful, seamless and transparent—and that the concept of machines being indistinguishable from a human is out of touch.
I sort of agree with that, but in certain applications I still think it would be potentially useful for a computer to be able to mimic human conversation. In fact, I believe that this has already been achieved to a limited extent. Albeit only so far in certain predefined situations. For example, a computer can mimic a human well enough to schedule an appointment at a hair salon (this was achieved two years ago) or book a hotel room, because that sort of conversation is not open-ended and includes only a finite number of basic concepts. But the voice sounds very natural, even adding waiting sounds like "umm, ..." to make it sound more like a person than a robot.

I think some people are seriously hoping to create real robot companions in the future, and to achieve that, the Turing test is a great barometer.

But of course, there are many other possible applications where mimicking a human is not necessary to the function. And nowadays it's probably more common to book an appointment via the internet rather than calling a receptionist to schedule one.
xouper
Posts: 11741
Joined: Fri Jun 11, 2004 4:52 am
Title: mere ghost of his former self

### Re: Turing test passed again and again treated as a prank.

Anaxagoras wrote: Tue Jan 05, 2021 4:08 am . . . But of course, there are many other possible applications where mimicking a human is not necessary to the function.
An example that comes immediately to mind is a driverless car.

I suspect no one wants to ride in a car driven by a computer whose driving ability is indistinguishable from a human driver.

A Turing Test would not be an adequate barometer of that computer's "intelligence".

As always, Your Mileage May Vary™.
robinson
Posts: 19621
Joined: Sat Aug 12, 2006 2:01 am
Title: Je suis devenu Français
Location: USA

### Re: Turing test passed again and again treated as a prank.

xouper wrote: Tue Jan 05, 2021 5:37 am I suspect no one wants to ride in a car driven by a computer whose driving ability is indistinguishable from a human driver.
Good lard

Nobody wants any computer that is just like a person. We want fucking computers that are better than a person.
Rob Lister
Posts: 23535
Joined: Sun Jul 18, 2004 7:15 pm
Title: Incipient toppler
Location: Swimming in Lake Ed

### Re: Turing test passed again and again treated as a prank.

Anaxagoras wrote: Tue Jan 05, 2021 4:08 am But the voice sounds very natural, even adding waiting sounds like "umm, ..." to make it sound more like a person than a robot.
That is particularly true with poorly implemented text chatting bots that insist on popping up on my screen to 'help' me. I find such "umm" additions very condescending.
xouper
Posts: 11741
Joined: Fri Jun 11, 2004 4:52 am
Title: mere ghost of his former self

### Re: Turing test passed again and again treated as a prank.

robinson wrote: Tue Jan 05, 2021 7:21 am Nobody wants any computer that is just like a person. We want fucking computers that are better than a person.
Like one of these? (perhaps nsfw)

Pyrrho
Posts: 33281
Joined: Sat Jun 05, 2004 2:17 am
Title: Man in Black
Location: Division 6

### Re: Turing test passed again and again treated as a prank.

https://openai.com/blog/dall-e/
DALL·E: Creating
Images from Text
We’ve trained a neural network called DALL·E that creates images from text captions for a wide range of concepts expressible in natural language.
Pyrrho
Posts: 33281
Joined: Sat Jun 05, 2004 2:17 am
Title: Man in Black
Location: Division 6

### Re: Turing test passed again and again treated as a prank.

https://www.scientificamerican.com/arti ... e-content/
Social media platforms, relying on thousands of human reviewers, are struggling to moderate the ever-increasing volume of harmful content. In 2019, it was reported that Facebook moderators are at risk of suffering from PTSD as a result of repeated exposure to such distressing content. Outsourcing this work to machine learning can help manage the rising volumes of harmful content, while limiting human exposure to it. Indeed, many tech giants have been incorporating algorithms into their content moderation for years.
robinson
Posts: 19621
Joined: Sat Aug 12, 2006 2:01 am
Title: Je suis devenu Français
Location: USA

### Re: Turing test passed again and again treated as a prank.

It is pretty obvious on Twitter

A human pushes a button on a name, an account, and then a robot watches the account, doing stupid shit because it’s not smart enough
robinson
Posts: 19621
Joined: Sat Aug 12, 2006 2:01 am
Title: Je suis devenu Français
Location: USA

### Re: Turing test passed again and again treated as a prank.

But that is another topic
xouper
Posts: 11741
Joined: Fri Jun 11, 2004 4:52 am
Title: mere ghost of his former self

### Re: Turing test passed again and again treated as a prank.

robinson wrote: Mon Feb 08, 2021 5:44 pm . . . a robot watches the account, doing stupid shit because it’s not smart enough
And that's how you pass the Turing test, when the stupid shit done by a computer is indistinguishable from the stupid shit a human does.
robinson
Posts: 19621
Joined: Sat Aug 12, 2006 2:01 am
Title: Je suis devenu Français
Location: USA

### Re: Turing test passed again and again treated as a prank.

Yep
robinson
Posts: 19621
Joined: Sat Aug 12, 2006 2:01 am
Title: Je suis devenu Français
Location: USA

### Re: Turing test passed again and again treated as a prank.

The difference of course, is the AI has no sense of the absurd, while a human at some point might have a moment of clarity

Which will quickly be resolved by using doublethink