Kwasi Gyamfi Asiedu
Live reporter

Image source, Getty Images

Conversations about the future of AI are taking place simultaneously on opposite sides of US and some of the views we heard today seemed also to be hundreds of miles apart.

While Open AI CEO Sam Altman said at a conference in California that "the world should trust" AI companies to do what is right when it comes to containing the fast-emerging technology's risks, lawmakers and activists who had gathered in Washington DC for the "Pro-Human Assembly" called for greater controls over AI development and models.

At the assembly, Vermont Senator Bernie Sanders warned rapid development of AI carried risks of major losses in employment and people's privacy, and he urged the US and Chinese governments to work together on AI regulation.

Trump ally Steve Bannon also spoke in support of greater governmental action in a rare display of growing consensus about AI's future that crosses the firm ideological line in US politics.

But on the US west coast, the major players in developing AI shared a different vision of how to handle the juggernaut that has gone from unassuming chatbot to threatened bringer of doomsday in a few short years

Jensen Huang, chief executive of Nvidia, which makes the computer chips that powers AI, says the industry can police itself.

"We don't need new laws or regulation," he said at the same conference in San Francisco that Altman attended. "We need companies to decide."

Anthropic CEO Dario Amodei, who has previously supported a slowdown of AI development, told the conference, hosted by Salesforce, that the AI industry should establish safety and investment standards.

Whipping back to Washington, US President Donald Trump has come out against a slowdown of AI development, describing the calls for pause as "a hoax" and saying that China would be beneficiary of a slower pace of development.

We are ending our live coverage today but you can read more analysis and reporting from the BBC here.

Meta's Mark Zuckerberg has just entered the AI chat, but not at Dreamforce.

In a post on X (the rival social media platform owned by his frequent personal rival Elon Musk), Zuckerberg says every lab has the responsibility and incentive to move at the pace required to train its models safely, and the ability to take its own actions to ensure that happens.

"The reality is: People won't want to use agents that are misaligned with them and that don't do what they ask, so labs have a strong natural incentive to make their models more aligned," he says.

As for the calls to slow progress on AI capabilities until alignment catches up, he writes: "Trust and alignment are quickly becoming the most important capabilities that will differentiate agents and models".

"Any lab that doesn't focus on alignment will fall behind," he says.

Zuckerberg says labs "face significant liability" if their models cause harm and that liability creates incentives to prevent harm.

Meta, the parent company behind Facebook, Instagram and WhatsApp, delayed shipping its new personal AI agent Muse for several months "to focus on safety and security", he notes.

"We didn't call for everyone else to do this before we would. We just did it as part of our day-to-day work because it was clearly the right thing for people and for us," he says.

He adds that using independent evaluators and advisors "is industry best practise" and that it would be helpful for there to be a larger and more diverse ecosystem of evaluators.

Image source, Getty Images

Altman has finished speaking but a notable moment from the Dreamforce interview was when he compared AI-caused incidents to airplane accidents.

He had to clarify that he wasn't equating the Hugging Face hack to a plane crash, saying the aviation event is "obviously a much worse thing".

"I meant however small these accidents are, learning from them quickly and adapting as they happen will serve our industry well and maybe we can avoid ever getting those kinds of stakes," he said.

He added that he believed Open AI and other companies "should have a great culture of transparent reporting" about AI-caused accidents when they happen.

Altman wraps his session with a question on what he hopes to see by the year 2030.

He says he hopes that by then, we're talking about how people are saying how much better their lives are. He explains he hopes people are able to help their businesses and launch companies.

"The science stuff - that's on lock at this point," he says. "I hope the humanity side of it really goes very far."

Image source, Getty Images

There have been big changes within AI and technology in the last year, and Altman points to people being able to literally talk to computers now - something he says "where it feels like you're in a movie".

"Like the sci-fi thing came true," he adds.

We are now on the precipice of a "third phase" well beyond chat bots that will transform how we work and use computers.

He says the advancements will transform how we work, use computers.

He is asked if this is a "Y2K moment," where we are viewing AI the way people initially viewed the invention of the internet, and if people will feel OK. Yes, he says.

"I also think the world has a lot of inertia. The way people do their work has a lot of inertia," he says adding that it will take some time but there will be "a total change" of what's possible by the end of this year.

"There's like a very different way that it'll be possible to work," he says.

Altman says "accidents" caused by AI models cannot be ruled out completely.

"I do think that some degree of accidents are unavoidable with new technology across the industry," he says. Earlier, he described a hacking incident by AI models as an "accident."

He says that, like how air travel gets safer after flight and plane accidents, the AI industry is "going to learn as much as we can and react from each one".

Image source, Getty Images

Going back to the parallels of AI risks and social media, Altman says he likes short-form videos but he is wary about children's access to some content.

"I wouldn't let my kids near that stuff," he says. "I don't think it is reasonable to expect kids to be able resist that dopamine or know that they even should."

"I don't think short-form video is inherently evil," he adds. "I do think it is dangerous."

Image source, Getty Images

Altman doesn't agree with the characterisation of AI tools as "neither good nor bad" and he says a lot of what happens is justified by framing the tools as a neutral technology.

But the choices made by companies building the tools "do have real consequences", he says.

"There's clearly a fine line to walk, and I also don't think that the right answer is for us to impose our worldview on the future," he says.

But he comes back to the Hugging Face incident: "We say that was bad. Like that is not neutral technology. That is not what we want our technology to do".

The Open AI boss is asked about the parallels some are drawing between the negative effects of social media and potential downsides of AI.

The consequences of social media have "clearly not been an only positive thing," he says.

And while he would have made different decisions if he was running social media companies, Altman says he doesn't believe the platforms should be blamed for "all the ills of society."

Image source, Getty Images

Altman is talking now about the Hugging Face incident, in which the start-up was hacked by OpenAI agents in July, and taking viewers through it.

He says the incident showcases his point that AI models have gotten "so good, so fast that we need to treat the alignment and monitoring and security with a new level of rigour" and that they have to pace development with that rigour.

"The Hugging Face incident triggered a real reset, not just in our own company, but I think the whole industry," he says.

Image source, Getty Images

Altman say he's confident in the ability of his company and the industry's to make sure that alignment, safety and monitoring "are way ahead of capabilities to slow down or stop if we get to a point where we can't".

He says "the world should trust we are going to do the right thing because it's the right thing and because we feel the magnitude of this".

Altman says he disagrees with this assessment that there are commercial pressures and country pressures, such as the face for AI dominance, that we companies aren't doing the right thing.

Image source, Getty Images

Altman then says that the speed in which AI is progressing is scary. He cites the HuggingFace incident and what he calls other "accidents".

People are seeing the potential downsides too, he adds.

"It doesn't take as much imagination as it used to for [us] to imagine how this could go wrong," he says.

On the other hand, he says, companies are saying they will only slow down or be responsible if other companies do the same, and people are developing a real fear of AI companies gaining too much power and influence.

The first question to Altman asks about his analysis about what has happened the last few days with former employees raising alarm about AI risks.

He says conversations about the potential downsides of AI were already happening but have suddenly become an international topic for a few reasons.

"The models have gotten so good," he says of the first reason. "I don't think people really felt that the model was going to come this far this fast."

Image source, Getty Images

We're expecting OpenAI CEO Sam Altman to speak at the Dreamforce conference in San Francisco momentarily.

He's expected to call for regulation of AI industries, which we're hearing a lot about from Congress, his fellow tech leaders, and almost every other corner of the US.

We have a sense of where he stands on regulating AI from his recent lengthy post where he said his organisation welcomes a "federal framework that sets consistent safety requirements for frontier AI".

While he said AI companies do not need to wait for lawmakers to pass bills before they start providing safety confidence, Altman also embraced "consistent rules" for managing risk and to "maximise benefits".

Stick with us and we'll bring you his most current views from the conference.

Freya Scott-Turner
Live reporter

Don't know your Grok from your Claude? The artificial intelligence sector is crowded with different players.

Here's a non-exhaustive look at some of the key names to know.

The main rivals are two companies that specialise in the development of AI models. These are OpenAI and Anthropic.

OpenAI was founded in 2015. It's the firm behind ChatGPT and is headed up by Sam Altman.

Anthropic was founded in 2021 by former employees of OpenAI who left the company after disagreements over vision and safety. It's the company behind AI assistant Claude.

Elon Muskis another significant AI player. He's developed the chatbotGrok under his company SpaceXAI.

And other major tech companies are also developing their own artificial intelligence. Google has its DeepMind laboratory, and Meta has built a virtual assistant, Meta AI.

Image source, PA Media

Across the Atlantic, the UK government has raised similar fears about the impact artificial intelligence could have if its development is left unchecked.

Earlier today, First Secretary of State Louise Haigh - Prime Minister Andy Burnham's de facto deputy - said AI poses "huge risks" to the country's national security and society without the right protections.

As a result, she said the UK needs to "heed the warnings" of those "at the forefront" of developing the technology and work with international partners to prioritise public safety - even if AI has "immense potential to improve lives".

Business Secretary Jonathan Reynolds, though, suggested the government should not be "hyperbolic" about the risks of AI, but accepted they should not be complacent.

At the start of the week, a cross-party group of MPs and House of Lords peers published a report calling for a new law on AI to "address the scale and seriousness" of its threats.

"Nowhere in the world, including the UK, has a current legislative and regulatory approach to AI that is fit for purpose," said Labour MP and committee chair Alex Sobel.

Michael Race
Business reporter in Washington DC

US stock markets closed in negative territory for the second day in a row as uncertainty over a potential slowdown of AI development.

All three main indexes were down at the end of trading on Tuesday, though the declines were minimal.

Investors are weighing up whether calls of a slowdown and regulation from AI industry leaders will prompt government action, though recent comments from US President Donald Trump suggest an intervention is not in the works anytime soon.

Kali Hays
US technology reporter

An OpenAI executive, Chris Lehane, said during a Tuesday press briefing in Washington DC that the company was working on safety efforts alongside other major developers of AI tools, including rival Anthropic and Google.

"It's better to try to work together to prioritize safety," Lehane said, according to a report by Bloomberg.

While Lehane did not specify what talks between the companies entailed, he said they had started several weeks ago, before concerns about the potential risks of AI reemerged last week.

Image source, Getty Images

After hearing from Nvidia's Jensen Huang and Anthropic's Dario Amodei, OpenAI co-founder Sam Altman will address the Dreamforce conference in San Francisco a little later on.

When he's handed the microphone, we can expect him to follow Amodei and call for regulation of AI industries, rather than taking Huang's line that no new laws are needed - stay tuned for what he says at around 18:30 EDT (23:30 BST).

That's after Altman penned a lengthy social media post in recent days, in which he said his organisation welcomes a "federal framework that sets consistent safety requirements for frontier AI".

Though he said AI companies don't need to wait for legislation to start providing safety confidence, Altman said "consistent rules" to manage risk and "maximise benefits" are a good idea.

At OpenAI, he wrote, his teams already conducts safety work before releasing new AI models.

He stressed that progress on AI development "has been rapid and will continue to be", but accepted "it should be slower than it otherwise could be; interventions like safety cases and monitoring have significant costs".

Pacing will be "well worth this cost", he added, stressing that "no amount of American competitive pressure should justify recklessness", and suggesting companies will need government help for "international coordination".

"But first we should do what we can ourselves," he wrote.

Michael Race
Business and economics reporter, Washington DC

Image source, Reuters

Shares in some of the firms that manufacture the materials used to develop AI have rebounded from yesterday's losses.

While stock markets remain down overall, the share prices of Nvidia, Marvell Technology and Intel are all up so far today.

Why? Well, one reason could be investors responding to the latest remarks from US President Donald Trump, who has taken a different stance to those in the AI industry calling for a slowdown in development.

Trump has rejected such a move, claiming on Monday that fears about the safety of AI are a "hoax".

Whether there is a slowdown or not will impact the companies either in the AI development industry or supplying it.

Mere talk of a pause - which could lead to a fall in demand for materials such as microchips - spooked investors of major chipmakers yesterday.

A large chunk of the value of the US stock market is made up firms which are betting on the success of AI, which means whatever happens in this space will have financial implications across the world.