Täällä seurataan AI-ripuligeneraattoreiden maailmanvalloitusta

[pigra senlaborulo]

Much like Weyland-Yutani linking arms with Skynet, Microsoft have teamed up with Inworld to bring the latter's "generative AI" tools to Xbox studios. This "multi-year co-development partnership" will "build a powerful toolkit that harnesses artificial intelligence to enrich the narrative and character creation elements of game development", according to Inworld CEO Ilya Gelfenbeyn. You might recognise Inworld as the tech used by that GTA 5 mod about AI cultists in which characters speak AI-generated dialogue, which Rockstar blocked from distribution over the summer. Suffice to say that Microsoft have grander plans for it.

Gelfenbeyn's announcement post (thanks Eurogamer) specifies two particular WIP tools, an "AI design copilot that assists and empowers game designers to explore more creative ideas, turning prompts into detailed scripts, dialogue trees, quests and more", and an "AI character runtime engine that can be integrated into the game client, enabling entirely new narratives with dynamically-generated stories, quests, and dialogue for players to experience."

Of the "copilot", the blog notes that it could "reduce the time and resource constraints during production to not only ship faster, but to also craft more expansive and immersive worlds and stories". It'll apparently "take the seeds of these creative ideas and transform them into detailed scripts complete with dialogue trees, quests, and more." The copilot is in closed beta right now and they'll be sharing more on 17th November, at the G-Star Korea trade show in Busan.

As for the "AI character runtime engine", it'll apply to expressions, dialogue and gestures, and will supposedly help games shift from "a scripted player-character dynamic to one that adapts to players' actions and choices in real-time" so as to avoid "repetitive dialogue and the inability to adapt to in-game actions".

Xbox general manager of gaming AI Haiyan Zhang adds in the Inworld post that the forthcoming toolkit will "empower creators in dialogue, story, and quest design", and will be available to developers of all sizes around the world. It'll be integrated with Microsoft's internal "AI" projects, including the Azure OpenAI Service and Microsoft Research's, er, research.

Microsoft are up to their necks in various AI projects and productions. The company announced that they had plunged billions into OpenAI this January, shortly after laying thousands of people off. The framing of the Inworld tool as a "copilot" and an "assistant", designed to cut out time rather than staff, reads like an attempt to reassure current Xbox developers that they aren't on the chopping block.

https://www.rockpapershotgun.com/micros ... r-creators

[pigra senlaborulo]

cnbc.com
'Overhyped' generative AI will get a 'cold shower' in 2024, analysts predict
Ryan Browne
6–8 minutes

An AI (Artificial Intelligence) sign is seen at the World Artificial Intelligence Conference (WAIC) in Shanghai, China July 6, 2023. REUTERS/Aly Song

An AI sign is seen at the World Artificial Intelligence Conference in Shanghai, July 6, 2023.

Aly Song | Reuters

The buzzy generative artificial intelligence space is due something of a reality check next year, an analyst firm predicted Tuesday, pointing to fading hype around the technology, the rising costs needed to run it, and growing calls for regulation as signs that the technology faces an impending slowdown.

In its annual roundup of top predictions for the future of the technology industry in 2024 and beyond, CCS Insight made several predictions about what lies ahead for AI, a technology that has led to countless headlines surrounding both its promise and pitfalls.

The main forecast CCS Insight has for 2024 is that generative AI “gets a cold shower in 2024” as the reality of the cost, risk and complexity involved “replaces the hype” surrounding the technology.

“The bottom line is, right now, everyone’s talking generative AI, Google, Amazon, Qualcomm, Meta,” Ben Wood, chief analyst at CCS Insight, told CNBC on a call ahead of the predictions report’s release.

“We are big advocates for AI, we think that it’s going to have a huge impact on the economy, we think it’s going to have big impacts on society at large, we think it’s great for productivity,” Wood said.

“But the hype around generative AI in 2023 has just been so immense, that we think it’s overhyped, and there’s lots of obstacles that need to get through to bring it to market.”

Generative AI models such as OpenAI’s ChatGPT, Google Bard, Anthropic’s Claude, and Synthesia rely on huge amounts of computing power to run the complex mathematical models that allow them to work out what responses to come up with to address user prompts.

Companies have to acquire high-powered chips to run AI applications. In the case of generative AI, it’s often advanced graphics processing units, or GPUs, designed by U.S. semiconductor giant Nvidia that large companies and small developers alike turn to to run their AI workloads.

Now, more and more companies, including Amazon, Google, Alibaba, Meta, and, reportedly, OpenAI, are designing their own specific AI chips to run those AI programs on.

“Just the cost of deploying and sustaining generative AI is immense,” Wood told CNBC.

“And it’s all very well for these massive companies to be doing it. But for many organizations, many developers, it’s just going to become too expensive.”
EU AI regulation faces obstacles

CCS Insight’s analysts also predict that AI regulation in the European Union — often the trendsetter when it comes to legislation on technology — will face obstacles.

The EU will still be the first to introduce specific regulation for AI — but this will likely be revised and redrawn “multiple times” due to the speed of AI advancement, they said.

“Legislation is not finalized until late 2024, leaving industry to take the initial steps at self-regulation,” Wood predicted.

Generative AI has generated huge amounts of buzz this year from technology enthusiasts, venture capitalists and boardrooms alike as people became captivated for its ability to produce new material in a humanlike way in response to text-based prompts.

The technology has been used to produce everything from song lyrics in the style of Taylor Swift to full-blown college essays.

While it shows huge promise in demonstrating AI’s potential, it has also prompted growing concern from government officials and the public that it has become too advanced and risks putting people out of jobs.

Several governments are calling for AI to become regulated.

In the European Union, work is underway to pass the AI Act, a landmark piece of regulation that would introduce a risk-based approach to AI — certain technologies, like live facial recognition, face being barred altogether.

In the case of large language model-based generative AI tools, like OpenAI’s ChatGPT, the developers of such models must submit them for independent reviews before releasing them to the wider public. This has stirred up controversy among the AI community, which views the plans as too restrictive.

The companies behind several major foundational AI models have come out saying that they welcome regulation, and that the technology should be open to scrutiny and guardrails. But their approaches to how to regulate AI have varied.

OpenAI’s CEO Sam Altman in June called for an independent government czar to deal with AI’s complexities and license the technology.

Google, on the other hand, said in comments submitted to the National Telecommunications and Information Administration that it would prefer a “multi-layered, multi-stakeholder approach to AI governance.”
AI content warnings

A search engine will soon add content warnings to alert users that material they are viewing from a certain web publisher is AI-generated rather than made by people, according to CCS Insight.

A slew of AI-generated news stories are being published every day, often littered with factual errors and misinformation.

According to NewsGuard, a rating system for news and information sites, there are 49 news websites with content that has been entirely generated by AI software.

CCS Insight predicts that such developments will spur an internet search company to add labels to material that is manufactured by AI — known in the industry as “watermarking” — much in the same way that social media firms introduced information labels to posts related to Covid-19 to combat misinformation about the virus.
AI crime doesn’t pay

Next year, CCS Insight predicts that arrests will start being made for people who commit AI-based identify fraud.

The company says that police will make their first arrest of a person who uses AI to impersonate someone — either through voice synthesis technology or some other kind of “deepfakes” — as early as 2024.

“Image generation and voice synthesis foundation models can be customized to impersonate a target using data posted publicly on social media, enabling the creation of cost-effective and realistic deepfakes,” said CCS Insight in its predictions list.

“Potential impacts are wide-ranging, including damage to personal and professional relationships, and fraud in banking, insurance and benefits.”

https://www.cnbc.com/2023/10/10/generat ... edict.html

[38911 BASIC BYTES FREE]

Taas päättyi yksi Mystery AI Hype Theater 3000, tallenne vielä hetken aikaa katsottavissa tuolta:

https://www.twitch.tv/dair_institute

Suosittelen kyllä näitä jutustelutuokioita, Emily M. Bender ja Alex Hanna ja välillä muutakin porukkaa juttelee ja käy läpi karmeimpia AI-hypeartikkeleita ja muuta aiheeseen liittyvää.

[pigra senlaborulo]

au.finance.yahoo.com
YouTube tests AI-generated comment summaries and a chatbot for videos
Will Shanklin
3 minutes

YouTube announced two new experimental generative AI features on Monday. YouTube Premium subscribers can soon try AI-generated comment summaries and a chatbot that answers your questions about what you’re watching. The features will be opt-in, so you won’t see them unless you’re a paid member who signs up for the experiments during their test periods.

The AI-powered summaries will organize comments into “easily digestible themes.” In a Mr. Beast video YouTube used as an example, the tool generated topics including “People love Bryan the bird,” “Lazarbeam should be in more videos,” “No submarine” and “More 7 day challenges.” You can tap on the topic to view the complete list of associated comments. The tool will only run “on a small number of videos in English” with large comment sections.

Screenshot from the YouTube mobile app with an AI-powered comment summary feature. It categorizes comments by topic.

Screenshot from the YouTube mobile app with an AI-powered comment summary feature. It categorizes comments by topic. (YouTube)

If you’re worried about YouTube’s summaries spiraling out of control the way the platform’s comment sections often do, the company says it won’t pull content from unpublished messages, those held for review, any containing blocked words or those from blocked users. Further, creators can use the tool to delete individual comments if they see problematic (or otherwise unwanted) discussions about their videos.

Meanwhile, YouTube’s conversational AI tool gives you a chatbot trained on whichever video you’re watching. Generated by large language models (LLMs), the assistant lets you “dive in deeper” by asking questions about the content and fishing for related recommendations. The company says the AI tool, which appears similar to chatting with Bard, draws on info from YouTube and the web, providing answers without interrupting playback. Eligible users can find it under a new “Ask” button in the YouTube app for Android.

Starting today, YouTube Premium subscribers can opt into the comment summarizer on YouTube’s experiments page. However, the company says you won’t see the “Topics” option for all videos. In addition, the conversational AI tool is only available now “to a small number of people on a subset of videos,” but YouTube Premium subscribers with Android devices will be able to sign up to try it in the coming weeks. The company warns the experimental features “may not always get it right,” a description that can equally apply to Google’s other AI experiments.
https://au.finance.yahoo.com/news/youtu ... 05231.html

[pigra senlaborulo]

An 82-year-old man was tricked out of thousands of dollars after scammers used artificial intelligence technology to pose as a Texas police officer over the phone, ABC 13's Erica Simon reported.

Jerry, identified only by his first name, told ABC 13 that he received a call on Oct. 21 from a stranger who introduced himself as Sgt. Matthew with the San Antonio Police Department. The caller told Jerry that Jerry's son-in-law, Michael Trueblood, was in jail for a serious accident that he caused. Another voice, which Jerry then thought belonged to his son-in-law, came on the line and asked Jerry to help pay for his release.

The scammers convinced Jerry to take out a total of $17,000 and hire a courier to deliver the cash from the assisted living facility where Jerry resided with his wife. Although Jerry wrote down the license plate numbers of the vehicles that drove off with his money, officers with the Sugar Land Police Department said leads to any linked courier services have turned up short.

"We really depended on that money. I'm going to have to get a job somewhere. H-E-B or something like that, to try to restore some of this money," Jerry said.


AI experts say there has been an uptick in similar schemes that rely on voice cloning devices over the past year. According to a FOX Business article by Eric Revell, advancements in technology have allowed criminals to better impersonate their targets' friends and families.

With audio samples that span as little as three to 10 seconds, scammers can realistically replicate a person's voice. These short clips can be taken from phone calls, social media, or other corners of the internet, Mike Scheumack, chief marketing and innovation officer with the identity protection company IDIQ, told Fox. "The goal of the scammer is to get you into fight or flight [mode] and to create urgency in your mind that your loved one is in some sort of trouble," he said. "So the best way to deal with those situations is to hang up and immediately call your loved one to verify if it’s them or not."

These cases also generally require multiple people to carry out, Scheumack added. "You have people that are researching on social media and gathering data about people. Those are not the same people that are going to plug in your voice. You have somebody else that’s going to clone the voice. You have somebody else that’s going to actually commit the act of calling. And you have somebody come to the victim’s house and pick up money if the scam is working."

As these operations proliferate, President Joe Biden recently signed an executive order calling for strict standards to manage AI safety and security. On top of directing software developers to release safety testing results and other critical information, he urged government agencies to set guidelines for improving cybersecurity and privacy, protecting U.S. citizens, and authenticating digital content.

"AI is all around us," Biden said, per an Associated Press report. "To realize the promise of AI and avoid the risk, we need to govern this technology."https://www.chron.com/news/houston-texa ... 469518.php

[pigra senlaborulo]

mikrobitti.fi
EY:ltä miljardi-investoinnit tekoälyyn
Suvi Korhonen
~1 minute

Tilintarkastus- ja konsulttiyhtiö EY on lanseerannut oman EY.ai-tekoälyalustansa. Sen kehittämiseen kului yhtiön mukaan 1,4 miljardia dollaria eli noin 1,3 miljardia euroa.

EY on tehnyt yhteistyötä Microsoftin kanssa, jonka GPT-3- ja GPT-4-pohjaisiin Azure OpenAI -toimintoihin EY:llä oli pääsy jo ennen varsinaista julkaisua. EY julkisti myös oman laajan kielimallinsa, EY.ai EYQ:n.

Alustan ekosysteemiyhteistyössä ovat lisäksi mukana muun muassa Dell Technologies, IBM, SAP, ServiceNow, Thomson Reuters ja UiPath.

Samalla tekoälyominaisuuksia on lisätty EY:n tuotteisiin, kuten analytiikka-alusta EY Fabriciin.

EY.ai Confidence Index mahdollistaa tekoälypohjaisen riskien ja tiedonhallinnan arvioinnin ja valvonnan. EY.ai Maturity Model vertailee yritystä kilpailijoihin ja kertoo, miten organisaatio sijoittuu alallaan. EY.ai Value Accelerator taas auttaa priorisoimaan hankkeita ja ratkaisuja kasvun ja strategian toteuttamisen näkökulmasta.

https://www.mikrobitti.fi/uutiset/mb/ab ... 7e8a1831dd

[pigra senlaborulo]

www.mikrobitti.fi
Androidin suosituin kuvasovellus sai tekoälyn: kieltäytyy muokkaamasta naamaasi
Samuli Leppälä

Googlen Kuvat-sovellukseen esitelty Magic Editor mahdollistaa kuvien tekoälyavusteisen muokkaamisen suoraan puhelimessa. Teknologiajätti on kuitenkin asettanut ominaisuudella tiukat rajoitukset.

Generatiivisen tekoälyn avustamana Magic Editor voi muokata kuvia tai tuoda mukaan kokonaan uusia elementtejä.

Google on asettanut mukaan joukon rajoitteita, jotka estävät Magic Editorin käyttämisen kyseenalaisiin tarkoituksiin, Android Authority kirjoittaa. Kuvat-sovelluksen koodia penkomalla verkkojulkaisu havaitsi, että Magic Editor kieltäytyy muokkaamasta henkilökortteja, kuitteja tai muita asiakirjoja.

Kiellettyjen asioiden listalla ovat myös kuvat henkilötiedoista, ihmiskasvoista ja ruumiinosista. Lisäksi isot kohteet, jotka vaatisivat paljon generoitua dataa, ovat kiellettyjä.

Rajoituksilla Google pyrkii epäilemättä rajoittamaan generatiivisen tekoälyn hyödyntämistä esimerkiksi rikollisiin tarkoituksiin. Toistaiseksi Magic Editor on saatavissa Pixel 8- ja Pixel 8 Pro -puhelimissa, mutta ominaisuus saattaa levitä muihin Android-puhelimiin tulevaisuudessa.
https://www.mikrobitti.fi/uutiset/mb/1c ... parit:17ff

[pigra senlaborulo]

www.mikrobitti.fi
Salaperäinen ”kännykäntappaja” paljastui: tekoäly rintataskuun, näyttö kämmenelle
Samuli Leppälä

Startup-yhtiö Humane on kerännyt paljon huomiota tekoälybuumin keskellä. Entisten Apple-työntekijöiden perustama yhtiö on jo pitemmän aikaa hehkuttanut kehittävänsä tuotetta, joka vapauttaa maailman älypuhelinten kynsistä.

Varsinainen tuote on AI Pin, joka on luotu yhteistyössä Microsoftin ja tekoäly-yhtiö OpenAI:n kanssa.

Paitaan tai takkiin kiinnitettävä laite tarjoaa tekoälyominaisuuksia helpottamaan käyttäjänsä arkea. Käytännössä kyse on mukana kulkevasta tekoälyassistentista, joka osaa hoitaa myös itsenäisenä laitteena puhelimen virkaa.

Aikaisemmin Ted-konferenssissa Humanen toinen perustaja Imran Chaudhri piti puheen, jossa hän esitteli yhtiön puettavaa teknologiaa. Tuolloin kävi muun muassa ilmi, että AI Pin kykenee projisoimaan kuvia käyttäjänsä kämmenelle.

Virallisesti tuote on määrä julkaista aivan pian, mutta The Verge kertoo saaneensa asiakirjoja koskien laitetta, sen ominaisuuksia ja kovaa hinnoittelua juuri ennen sen julkistusta.

LUE MYÖS:

Näiden tietojen mukaan laitteen hintalappu olisi kirvelevät 699 dollaria. Tämän lisäksi laitteen kanssa on maksettava peräti 24 dollarin kuukausimaksua, johon kuuluu mobiiliyhteys sekä pääsy Microsoftin ja OpenAI:n kielimalleihin.

Varsinainen AI Pin kiinnittyy paikoilleen magneetilla, jonka toinen puolisko toimii myös virtapankkina. Tarpeen vaatiessa käyttäjä voi yksinkertaisesti vaihtaa magneetin toisen puoliskon uuteen ja jatkaa laitteen käyttöä. AI Pinin mukana toimitetaan kaksi ”battery boosteria”.

Laitteessa on sisäänrakennettuna kamera sekä syvyys- ja liikenneanturit, joilla se tarkkailee ympäristöään. Kameralle voi esitellä erilaisia ympäröivän maailman objekteja, mutta laite ei jatkuvasti tallenna sen ympäristöä. Mukana on merkkivalo, joka kertoo kameran olevan käytössä.

Laskentatehosta vastaa Qualcomm Snapdragon -järjestelmäpiiri, jonka varsinaista mallia ei ole paljastettu. On myös mahdollista, että osa laskennasta suoritetaan pilvessä.

AI Pin käyttää omaa kaiutinta, mutta laitteessa on tuki myös bluetooth-kuulokkeille. Äänikomentojen lisäksi laitteelle voi jollain tapaa antaa myös elekomentoja. Ted-esityksessä näytetty kuvien projisointi pinnoille on myös mukana.

Asiakirjoista käy ilmi, että laite on todellakin suunniteltu itsenäiseksi kokonaisuudeksi, joka on käytettävissä älypuhelimesta erillisinä laitteena. AI Pinin käyttöjärjestelmän nimi on Cosmos, mutta laitetta on käytettävä Humane.center-keskuksen kautta. Toistaiseksi on epäselvää, mitä tämä tarkalleen ottaen tarkoittaa.

Varsinainen laite on määrä paljastaa virallisesti jo tänään 9. marraskuuta.
https://www.mikrobitti.fi/uutiset/mb/f2 ... 3597d2f37a

[☽☽☽]

Toistaiseksi on epäselvää, mitä tämä tarkalleen ottaen tarkoittaa.

[Spandau Mullet]

Tuolloin kävi muun muassa ilmi, että AI Pin kykenee projisoimaan kuvia käyttäjänsä kämmenelle.

Kuulostaa tutulta kusetukselta, now with AI

[pigra senlaborulo]

mikrobitti.fi
Legendaarinen Paint-ohjelma sai tekoälyn – kuinkas muutenkaan?
Olli Vänskä
~2 minutes

Muistatko Paintin?

Microsoftin piirto-ohjelma julkaistiin 38 vuotta sitten. Softa on pysynyt lähinnä kuriositeettina ja usein lähinnä viimeisenä vaihtoehtona piirtämiseen tai kuvien muokkaamiseen – sen ominaisuudet kun ovat olleet varsin rajoittuneita. Viime vuosikymmeninä sitä onkin käytetty ensisijaisesti meemimielessä.

Nyt Paint on kuitenkin saanut läjän uusia kikkoja – ja lisää on tulossa.

Mikrobitti kertoi syyskuussa, että versiossa 11.2308.18.0 Paint sai muun muassa kuvatasot (layers) ja sen mukana myös paremman tuen läpinäkyvyydelle.

Windowsin blogissa kerrottiin kuitenkin jo lokakuussa uusista lisäominaisuuksista, jotka tulivat käyttöön ensimmäisinä luonnollisesti kokeellisen Windows Insider -ohjelman osanottajille.

Versio 11.2309.20.0:n uudistuksiin kuuluu Paint Cocreator -ominaisuus, jonka myötä muokattavaan kuvaan voi lisätä sisältöä Dall-E 3:n avulla. Oheisessa tviitissä kuvaan luodaan tekstikäskyllä esimerkinomaisesti ”viidakkopolku ja puiden välistä kimaltavia auringonsäteitä”.

Dall-E 3:n avulla voi luonnollisesti myös luoda hahmoja ja kokonaisia kuvia, ei pelkästään muokattavien kuvien taustoja.

Microsoft kertoo blogissaan, että Cocreatorin käyttöä rajoitetaan tietyiltä osin, ”jotta voidaan estää luomasta kuvia, jotka voivat olla haitallisia, loukkaavia tai epäsäädyllisiä”.

Paintia on karsastettu laajasti, mutta pieniin muokkauksiin se on nykyiselläänkin aivan toimiva sovellus. Muun muassa tämän artikkelin Microsoftilta lainattu webp-kuva muunnettiin sen avulla käden käänteessä arkisemmaksi jpegiksi, jota CMS-järjestelmätkin ymmärtävät paremmin.

https://www.mikrobitti.fi/uutiset/mb/8d ... ffbc629ee4

[☽☽☽]

näpit irti paintista veijo microsoft

[pigra senlaborulo]

nature.com
‘ChatGPT detector’ catches AI-generated papers with unprecedented accuracy
Prillaman, McKenzie
5–7 minutes

A machine-learning tool can easily spot when chemistry papers are written using the chatbot ChatGPT, according to a study published on 6 November in Cell Reports Physical Science1. The specialized classifier, which outperformed two existing artificial intelligence (AI) detectors, could help academic publishers to identify papers created by AI text generators.

“Most of the field of text analysis wants a really general detector that will work on anything,” says co-author Heather Desaire, a chemist at the University of Kansas in Lawrence. But by making a tool that focuses on a particular type of paper, “we were really going after accuracy”.

The findings suggest that efforts to develop AI detectors could be boosted by tailoring software to specific types of writing, Desaire says. “If you can build something quickly and easily, then it’s not that hard to build something for different domains.”
The elements of style

Desaire and her colleagues first described their ChatGPT detector in June, when they applied it to Perspective articles from the journal Science2. Using machine learning, the detector examines 20 features of writing style, including variation in sentence lengths, and the frequency of certain words and punctuation marks, to determine whether an academic scientist or ChatGPT wrote a piece of text. The findings show that “you could use a small set of features to get a high level of accuracy”, Desaire says.

How ChatGPT and other AI tools could disrupt scientific publishing

In the latest study, the detector was trained on the introductory sections of papers from ten chemistry journals published by the American Chemical Society (ACS). The team chose the introduction because this section of a paper is fairly easy for ChatGPT to write if it has access to background literature, Desaire says. The researchers trained their tool on 100 published introductions to serve as human-written text, and then asked ChatGPT-3.5 to write 200 introductions in ACS journal style. For 100 of these, the tool was provided with the papers’ titles, and for the other 100, it was given their abstracts.

When tested on introductions written by people and those generated by AI from the same journals, the tool identified ChatGPT-3.5-written sections based on titles with 100% accuracy. For the ChatGPT-generated introductions based on abstracts, the accuracy was slightly lower, at 98%. The tool worked just as well with text written by ChatGPT-4, the latest version of the chatbot. By contrast, the AI detector ZeroGPT identified AI-written introductions with an accuracy of only about 35–65%, depending on the version of ChatGPT used and whether the introduction had been generated from the title or the abstract of the paper. A text-classifier tool produced by OpenAI, the maker of ChatGPT, also performed poorly — it was able to spot AI-written introductions with an accuracy of around 10–55%.

The new ChatGPT catcher even performed well with introductions from journals it wasn’t trained on, and it caught AI text that was created from a variety of prompts, including one aimed to confuse AI detectors. However, the system is highly specialized for scientific journal articles. When presented with real articles from university newspapers, it failed to recognize them as being written by humans.
Wider issues

What the authors are doing is “something fascinating”, says Debora Weber-Wulff, a computer scientist who studies academic plagiarism at the HTW Berlin University of Applied Sciences. Many existing tools try to determine authorship by searching for the predictive text patterns of AI-generated writing rather than by looking at features of writing style, she says. “I’d never thought of using stylometrics on ChatGPT.”

But Weber-Wulff points out that there are other issues driving the use of ChatGPT in academia. Many researchers are under pressure to quickly churn out papers, she notes, or they might not see the process of writing a paper as an important part of science. AI-detection tools will not address these issues, and should not be seen as “a magic software solution to a social problem”.

https://www.nature.com/articles/d41586-023-03479-4

[pigra senlaborulo]

What just happened? In case you missed it (we did), a knock-off version of Angry Birds called Angry Pumpkins launched on Halloween. It plays just like the classic game from 2010, except it has you shooting pumpkins at boxes, wooden planks, stone pillars, bones, skulls, and ghoulies.

Angry Pumpkins is not as advanced as its inspiration. It only has one level with seemingly no win condition, but it does have a crude editor allowing you to create custom layouts, sort of, maybe, if you are very, very patient. Okay, okay. The game sucks and is not very fun. You can judge for yourself on the developer's website.

But that's not the point.

What makes Angry Pumpkins newsworthy is that it was created entirely with natural language using ChatGPT to write its 600 lines of code, Midjourney to draw the objects and background, and Dall-E to generate the title screen. No manual coding was required, and little to no programming knowledge was needed to build the playable game.

Angry Pumpkins creator Javi Lopez has been tinkering with these AI tools for some time and has developed a keen sense of how to get the most out of them with English prompts. He even has tutorials on how to work with AI agents. Despite his better-than-average knowledge, he is still amazed he could create a playable game using only AI.

Midjourney, DALL*E 3 and GPT-4 have opened a world of endless possibilities.

I just coded "Angry Pumpkins ðÂÂÂ" (any resemblance is purely coincidental ðÂÂÂ) using GPT-4 for all the coding and Midjourney / DALLE for the graphics.

Here are the prompts and the process I followed: pic.twitter.com/st3OEhVVtK
– Javi Lopez â©ï¸Â (@javilopen) October 31, 2023

"I have to admit, I'm genuinely blown away," Lopez said on X. "Honestly, I never thought this would be possible. I truly believe we're living in a historic moment that we've only seen in sci-fi movies up until now."

A few of us at TechSpot have experimented with ChatGPT to code tasks in various programming languages. While it rarely produces completely functional error-free code on the first try, we've gotten it to self-correct and generate workable snippets for simple tasks. However, creating an interactive game is on a whole other level, even as crude as Angry Pumpkins is.

That said, designing the game was not as easy as pie. Lopez said that creating the graphics was the simplest part. He has been generating AI images for over a year and a half, and getting Dall-E or Midjourney to spit out what he wants was simply a matter of asking for it in the way the AI best understands. However, coding was a bear.

"Although the game is just 600 lines, of which I haven't written ANY, this was the most challenging part," Lopez said.

Getting ChatGPT-4 to produce workable code took patience and many, many iterations. He started with a simple prompt.

"Can we now create a simple game using matter.js and p5.js in the style of 'Angry Birds?' Just launch a ball with angle and force using the mouse and hit some stacked boxes with 2D physics."

Through a long back-and-forth process of testing and re-prompting, Lopez eventually got to a working, error-free game/demo. Never did he correct the code himself. If he got an error, he told ChatGPT what it was and commanded it to fix it. Sometimes, his prompts were humorously critical of the AI.

"Now, make the monsters circular, and be very careful: apply the same technique that already exists for the rectangular ones regarding scaling and collision area, and don't mess it up like before."

Despite the game's shortcomings, Angry Pumpkins is a great early example of how independent producers and even mainstream studios could leverage AI to save development costs. It also opens the door for virtually anybody with imagination and patience to create a game without any knowledge or experience in game development.

However, it's also a little scary as AI implementation in studios will most assuredly lead publishers to cut human staff, replacing them with AI. We aren't there yet, but the writing is on the wall.

"These new work processes, where we can create anything using just natural language, are going to change the world as we know it," Lopez opined. "It's such a massive tidal wave that those who don't see it coming will be hit hard. So... let's start riding the wave!"

https://www.techspot.com/images2/news/b ... _1100.webp

[pigra senlaborulo]

maksumuurin takana tämmönen. tuleekohan sieltä joku AI-potti

Suojelupoliisissa käynnistyy EU:n osittain rahoittama tekoälyä soveltava järjestelmähanke. Samalla rahoitusta myönnettiin useille poliisiorganisaatioille ja monessa hankkeessa oli kyse it-järjestelmien kehittämisestä.

https://www.mikrobitti.fi/uutiset/mb/60 ... 1345f7d8d3