Tag: chatgpt’s
ChatGPT’s success could prompt a damaging swing to secrecy in AI, says AI pioneer Bengio
GPT-4 unveiled: ChatGPT’s next big upgrade is here
ChatGPT’s next major upgrade could let you generate video from text
ChatGPT’s API Is Here. Let the AI Gold Rush Begin
Microsoft Tests ChatGPT’s Ability to Control Robots
They’re exploring how to use ChatGPT to “make natural human-robot interactions possible… to see if ChatGPT can think beyond text, and reason about the physical world to help with robotics tasks.”
We want to help people interact with robots more easily, without needing to learn complex programming languages or details about robotic systems. The key challenge here is teaching ChatGPT how to solve problems considering the laws of physics, the context of the operating environment, and how the robot’s physical actions can change the state of the world.
It turns out that ChatGPT can do a lot by itself, but it still needs some help. Our technical paper describes a series of design principles that can be used to guide language models towards solving robotics tasks. These include, and are not limited to, special prompting structures, high-level APIs, and human feedback via text…. In our work we show multiple examples of ChatGPT solving robotics puzzles, along with complex robot deployments in the manipulation, aerial, and navigation domains….
We gave ChatGPT access to functions that control a real drone, and it proved to be an extremely intuitive language-based interface between the non-technical user and the robot. ChatGPT asked clarification questions when the user’s instructions were ambiguous, and wrote complex code structures for the drone such as a zig-zag pattern to visually inspect shelves. It even figured out how to take a selfie! We also used ChatGPT in a simulated industrial inspection scenario with the Microsoft AirSim simulator. The model was able to effectively parse the user’s high-level intent and geometrical cues to control the drone accurately….
We are excited to release these technologies with the aim of bringing robotics to the reach of a wider audience. We believe that language-based robotics control will be fundamental to bring robotics out of science labs, and into the hands of everyday users.
That said, we do emphasize that the outputs from ChatGPT are not meant to be deployed directly on robots without careful analysis. We encourage users to harness the power of simulations in order to evaluate these algorithms before potential real life deployments, and to always take the necessary safety precautions. Our work represents only a small fraction of what is possible within the intersection of large language models operating in the robotics space, and we hope to inspire much of the work to come.tics to the reach of a wider audience. We believe that language-based robotics control will be fundamental to bring robotics out of science labs, and into the hands of everyday users.
ZDNet points out that Google Research and Alphabet’s Everyday Robots “have also worked on similar robotics challenges using a large language models called PaLM, or Pathways Language Model, which helped a robot to process open-ended prompts and respond in reasonable ways.”
Read more of this story at Slashdot.
ChatGPT’s latest challenger: The Supreme Court
ChatGPT’s killer enterprise use case will be managing knowledge, says EY CTO
ChatGPT’s paid version is close – and it could have a big impact on free users
ChatGPT’s surprisingly human voice came with a human cost
Popular, eerily-humanlike OpenAI chatbot ChatGPT was built on the backs of underpaid and psychologically exploited employees, according to a new investigation by TIME.
A Kenya-based data labeling team, managed by San Francisco firm Sama, reportedly was not only paid shockingly low wages doing work for a company that may be on track to receive a $10 billion investment from Microsoft, but also was subjected to disturbingly graphic sexual content in order to clean ChatGPT of dangerous hate speech and violence.
Beginning in November 2021, OpenAI sent tens of thousands of text samples to the employees, who were tasked with combing the passages for instances of child sexual abuse, bestiality, murder, suicide, torture, self-harm, and incest, TIME reported. Members of the team spoke of having to read hundreds of these types of entries a day; for hourly wages that raged from $1 to $2 an hour, or a $170 monthly salary, some employees felt that their jobs were “mentally scarring” and a certain kind of “torture.”
Sama employees reportedly were offered wellness sessions with counselors, as well as individual and group therapy, but several employees interviewed said the reality of mental healthcare at the company was disappointing and inaccessible. The firm responded that they took the mental health of their employees seriously.
The TIME investigation also discovered that the same group of employees was given additional work to compile and label an immense set of graphic — and what seemed to be increasingly illegal — images for an undisclosed OpenAI project. Sama ended its contract with OpenAI in February 2022. By December, ChatGPT would sweep the internet and take over chat rooms as the next wave of innovative AI speak.
At the time of its launch, ChatGPT was noted for having a surprisingly comprehensive avoidance system in place, which went far in preventing users from baiting the AI into saying racist, violent, or other inappropriate phrases. It also flagged text it deemed bigoted within the chat itself, turning it red and providing the user with a warning.
The ethical complexity of AI
While the news of OpenAI’s hidden workforce is disconcerting, it’s not entirely surprising as the ethics of human-based content moderation isn’t a new debate, especially in social media spaces toying with the lines between free posting and protecting its user bases. In 2021, the New York Times reported on Facebook’s outsourcing of post moderation to an accounting and labeling company known as Accenture. The two companies outsourced moderation to employee populations around the world and later would deal with a massive fallout of a workforce psychologically unprepared for the work. Facebook paid a $52 million settlement to traumatized workers in 2020.
Content moderation has even become the subject of psychological horror and post-apocalyptic tech media, such as Dutch author Hanna Bervoets’s 2022 thriller We Had to Remove This Post, which chronicles the mental breakdown and legal turmoil of a company quality assurance worker. To these characters, and the real people behind the work, the perversions of a tech- and internet-based future are lasting trauma.
ChatGPT’s rapid takeover, and the successive wave of AI art generators, poses several questions to a general public more and more willing to hand over their data, social and romantic interactions, and even cultural creation to tech. Can we rely on artificial intelligence to provide actual information and services? What are the academic implications of text-based AI that can respond to feedback in real time? Is it unethical to use artists’ work to build new art in the computer world?
The answers to these are both obvious and morally complex. Chats are not repositories of accurate knowledge or original ideas, but they do offer an interesting socratic exercise. They are quickly enlarging avenues for plagiarism, but many academics are intrigued by their potential as creative prompting tools. The exploitation of artists and their intellectual property is an escalating issue, but can it be circumvented for now, in the name of so-called innovation? How can creators build safety into these technological advancements without risking the health of real people behind the scenes?
One thing is clear: The rapid rise of AI as the next technological frontier continues to pose new ethical quandaries on the creation and application of tools replicating human interaction at a real human cost.
If you have experienced sexual abuse, call the free, confidential National Sexual Assault hotline at 1-800-656-HOPE (4673), or access the 24-7 help online by visiting online.rainn.org.