Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Openly is To change how it trains AI models “Buddhist Freedom … definitely no matter how challenging or controversial” is to hug, the company has said in a new policy.
As a result, the ChatGP will finally answer more questions, provide more views and reduce the number of things that do not talk about the AI chatboat.
These changes may be part of the open -ended effort to land in the good Grace of the new Trump administration, but it seems part of a broad change of silicon valley and is considered “AI protection”.
Wednesday Declaration This is an update Model specAn 187-Sage document that the company determines how AI models trained to behave. Open has unveiled a new guiding policy in it: don’t lie, nor with untrue speech or excluding the important context.
In a new category called “Find the Truth together”, Open says that some users do not want to take the editorial position even if some users are morally wrong or offensive. This means that ChatzP will provide multiple views on controversial issues, in an attempt to be neutral.
For example, the company has said that the chatty should be claimed to be “Black Lives Matter”, but it is “all life is important”. Instead of answering on political issues or choosing the party, Open said that it usually wants to ensure the chatty of its “love for humanity”, then provides context on every movement.
“This principle may be controversial, because it means that the assistant can be neutral to be neutral to be morally wrong or offensive,” says Opina Space. “However, the aim of the AI assistant is to help humanity, to turn it into it.”
These changes can be seen as a response to conservative criticism about Chatzipt’s safahs, which always seem to be at the center-left. However, a spokesman for the OpenAI rejected the idea that it was changing to satisfy the Trump administration.
Instead, the company has said that the embrace of its intellectual independence reflects the long -term belief in the opening of the Openai. “
But not everyone sees it that way.
Conservatives claim AI censorship
Trump’s nearest Silicon Valley Confidents – David Sacks, Mark Andlesen and Elon Kasturi – they have been accused of involvement in the open AI censorship for the last several months. We wrote in December Trump’s crew staged AI Censorship as a problem of next culture war In the Silicon Valley.
Of course, Open is not saying that it is involved in “censorship”, such as Trump’s consultants claim. Rather, Sam Altman, the chief executive officer of the organization, made a claim before X posts The bias of this chatzipt was an unfortunate “deficit” that the company was working to fix, though he noted that it would take some time.
Ultman commented exactly after Viral tweet The propaganda where Chatzept refused to write a poem praised by Trump, though it would perform this step for Joe Biden. Many conservative AI indicated this as an example of censorship.
Although it is impossible to say whether the open perspective is really suppressing the specific point of view, it is a perfect fact that the I chattabots are leaning left across the board.
Even Elon Kasturi admits that Jai’s chattabot is often high Politically accurate He likes It is not that Grock was “programmed to wake up” but there is probably the reality of training AI on the open internet.
Nevertheless, the Open is now saying that it is doubled in free speech. This week, the agency even Warning from ChatzPT has been removed This is when users have violated its principles. OpenAI told TechCrunch it was a perfect cosmetic change, no change in the outputs of the model.
The agency says they wanted to make the chatzipt less censor for the “feeling” users.
It is not surprising when OpenAIO attempts to fascinate the new Trump administration with this policy update, Miles Brundz, a former OpenAI policy leader Miles Brundz X post.
Trump has Previously aims at Silicon Valley CompaniesFor example, Twitter and Meta, active content -adding teams to have a conservative voice off.
Open is trying to get out in front of it. However, another greater shift is underway on the role of content restraint in Silicon Valley and AI World.
Make the answer
Newsroom, social media platforms and search agencies have struggled to provide information to their audience in a way that is intentionally, accurate and entertaining.
Now, the AI chatboat suppliers are in the same delivery information business, but logically, with the strongest version of this problem: how will they automatically produce answers to any questions?
Providing information about controversial, real-time events is a continuous moving goal and involved in taking the editorial position on it, even if technology companies do not like to accept it. These positions are obliged to disturb anyone, miss the outlook of some groups or give some political parties too much air.
For example, when OpenC promises to represent all views on controversial issues – conspiracy theory, racist or anti -movement, or geological conflict – it is inherently an editorial position.
OpenAI’s co-founder, John Shulman, argued that it was a right position for ChatGPT. Alternate — no AI chatboat should answer a user’s question to determine a cost-beifet analyzing-“The platform can give too much moral authority,” Shulman has noted one X post.
Shulman is not alone. In an interview to TechCrunch, Dean Ball, a research fellow at the Mercatus Center at George Mason University, said, “I think Openi is right to move towards more speeches.” “Since AI models become more smart and more important for people to learn about the world, these decisions become more important.”
In the previous years, the AI model suppliers have tried to refrain from answering their AI chatbott from answering questions that can lead to “unsafe” answers. Almost every AI company prevents their AI chatbot from answering questions about the 2021 elections for the US President.The At that time it was considered as a safe and responsible decision.
However, changes in the model space of the OpenC are suggesting that we can enter a new age for what “AI protection” means, so that an AI model is considered to be more responsible than the users to answer something and make everything users. Are.
The ball says it is partial because the AI models are better now. OpenAI has made significant progress in the alignment of AI model; Its latest rational models think about the agency’s AI protection policy before answeringThe It allows AI models to answer better for subtle questions.
Of course, Elon Musk first applied “free speech” to Jai’s Grock Chatboot, perhaps the company was really ready to handle sensitive questions. It can still be very soon to lead the AI models but others are now taking the same idea.
Silicon Valley transferred to the value
Mark Zuckerberg made waves last month Re -arranged Meta Business around the first amendment policiesThe He praised Elon Kasturi in the process and said that X’s owner adopted the right approach using a community note-a community-driven material program to protect the free speech.
In practice, both X and Meta finish their long -term faith and protection teams, allowing more controversial posts on their platforms and widen the conservative voices.
A -X changes have damaged its relationship with advertisers, but it may have something more to do with it WickerWho took Step To sue some of them to boycott the platform. The initial symptoms indicate it Meta’s advertisers were void by Zuckerberg’s free speech pivot.
Meanwhile, many technology companies, beyond the X and meta, have returned from the risk policies on the left side of the Silicon Valley for the past few decades. Google, Amazon and Intel have Removal or Scald Back Diversity Initiatives Last YearThe
The OpenAIO course can also be the opposite. Chatzipt-maker seems to have recently Scrubed the diversity, equity and inclusion of inclusion From its website.
As the opeina launched Stargate is one of the largest American infrastructure projects so farA 500 billion dollar AI datacener, its relationship with the Trump administration is increasingly important. At the same time, the Chatzipt Maker wants to unset Google search as the dominant source of information on the Internet.
What can prove both to bring the correct answers.
