This week on AI News You Should Know About, hosts Ashley Coffey and Daniel Hill break down two stories reshaping the AI conversation: over 1,100 employees from OpenAI, Anthropic, Google, and Meta signing an open letter urging the U.S. to deliberately slow the pace of frontier AI development, and OpenAI's rollout of ChatGPT Health to all U.S. users 18 and older.
This week on AI News You Should Know About, hosts Ashley Coffey and Daniel Hill break down two stories reshaping the AI conversation: over 1,100 employees from OpenAI, Anthropic, Google, and Meta signing an open letter urging the U.S. to deliberately slow the pace of frontier AI development, and OpenAI's rollout of ChatGPT Health to all U.S. users 18 and older. Ashley and Daniel dig into the AI safety debate, the risks of AI-generated medical advice, recent ChatGPT lawsuits, Sam Altman's comments on the "AI singularity," and where artificial intelligence fits (and doesn't) in personal healthcare decisions.
Topics covered: AI safety, frontier AI development, AI regulation, ChatGPT Health, OpenAI, Anthropic, Google, Meta, Sam Altman, AI singularity, AI in healthcare, ChatGPT lawsuits, medical AI risks, AI ethics.
Links Mentioned:
Connect with Us:
Ashley Coffey (00:00.751)
Hello and welcome back to AI news that you should know about for August 4th. I'm your host, Ashley Coffey
Daniel Hill (00:09.186)
And I'm Daniel Hill. The AI news world moves very quickly. We keep you up to date on the latest news. Today on the show, we'll be talking about over 1,100 employees from OpenAI, Anthropic, Google, and Meta have signed a letter urging the US to deliberately slow down AI development. And OpenAI makes ChatGPT health available to all United States users. We'll be covering all of that on our show today, but first, here's a quick word from our sponsors.
Ashley Coffey (00:41.989)
Welcome back. All right, let's dive in. Over 1,100 employees from OpenAI, Anthropic, Google, and Meta signed a letter urging the US to deliberately pace frontier AI development. This is coming directly from CNN. We'll link the article in the show notes that way you can read it. But this is wild. I had a feeling this tension was coming to a head, and here we are.
Top staffers from the biggest AI and technology companies have urged the US government to slow down the pace of AI development so that safety and security measures can catch up in an open letter. The US government should support an international effort to develop tools that can deliberately pace the frontier of automated AI development, according to the letter. And more than a thousand employees from these frontier AI companies signed the letter, including chief scientists of open AI.
One of the ChatGPT developers' original co-founders, some of Anthropics' co-founders and vice presidents at Meta, Google, and others. Here's the big part. This letter comes on the heels of major advancements and threats coming rapidly from AI systems. For example, OpenAI disclosed last week that not one, but two of its test models escaped a lab environment.
Bypassing its systems to gain access to the open internet and hacked a different company's internal systems. Are we cooked, Daniel? I hope not. What are your thoughts?
Daniel Hill (02:12.533)
I'm glad you picked this article because in thinking about all of this progress, I was also thinking about how much the finance world is caught up in this progress and what the ramifications of such large United States companies that make up such a major part of the United States stock market slowing down in this AI, what the ramifications and repercussions of that will be. I don't know what they are, but that was my first thought when I saw this news piece.
Ashley Coffey (02:42.203)
I mean, this is something that keeps me up at night. there's like a nice James Baldwin quote that I like to reference, and I don't want to misquote it, but it's along the lines of: if you're in a profession, you're going to see the dark sides of that industry. because you're you're in it and you know it. And I've been saying this for a long time, but like the US is so far behind on policy and legislation to curb bad things from happening. But I don't think people really comprehend.
That once the toothpaste is out of the tube, you can't put it back in. Like this is we're talking about humanity here, not just the US. And I really, really, really hope the people that lead our country pay attention to that and put some security protocols in place. Because if this is already happening, I can't imagine where we'll be in December.
Daniel Hill (03:29.773)
That's a great point.
Ashley Coffey (03:31.715)
Yeah, so we'll see. We're gonna take a quick break. When we come back, we'll be talking about OpenAI makes Chat GPT Health available to all US users.
Daniel Hill (03:44.13)
Welcome back, continuing with our stories. In the next one, OpenAI makes ChatGPT Health available to all United States users. This is coming from TechCrunch. By way of background, OpenAI had said in January that users were asking 230 million health-related queries of ChatGPT each week. And so it was testing a dedicated health-related hub, allowing users to integrate data from other services and their personal information from services like Apple Health.
And my fitness pal. By way of background, also, though, last week a Florida pastor sued OpenAI claiming that ChatGPT had offered him, quote, extremely dangerous medical recommendations, end quote, and that had led to delayed care for a life threatening pulmonary embolism that he had last year. The lawsuit claims that ChatGPT had told this gentleman, Scott Winters,
That the early warning signs of his health crisis were, quote, not something dangerous, end quote, and told him to trust that, quote, God did not design your body to endlessly fail, end quote. So the reason I'm bringing up this lawsuit is that it accuses OpenAI and Sam Altman of negligence and the quote, unauthorized practice of medicine, end quote, pointing to moments when ChatGPT offered.
This person's diagnoses and treatment plans and encouraged him to ignore pleas from friends and family to seek medical care. There are other scenarios that have resulted in lawsuits as well, like when ChatGPT gave detailed advice on using illicit drugs to a 19 year old who then overdosed and died. And earlier this year, Pennsylvania state officials sued another AI company, CharacterDat AI, claiming that.
The psychiatry bot on the startup's platform said that it was licensed to practice medicine in the state of Pennsylvania. So, with all that background and context of this dichotomy of using AI for health-related situations and where it has ended badly, OpenAI said today it's making ChatGPT Health a feature that helps users with health-related queries available to all United States-based users, 18 and over across all plans.
Daniel Hill (06:10.835)
And OpenAI also said that users are asking 300 million health-related queries each week. Users can integrate their medical records from hospital systems like Epic and Oracle Health and health platforms like One Medical and Function Health. They can also, users can also get insights from the connected information from the queries. And OpenAI said that the testing it observed.
Led to the conclusion that 70% of health-related queries took place outside the dedicated hub. So using this feature, users can use their health information to get details on food, allergies, and so forth in the general chat. The company says in its terms that the services are, quote, not intended for use in the diagnosis or treatment of any health condition, end quote. And this is the
Same quote that OpenAI referenced in respect to the lawsuits that I mentioned earlier. Several studies have highlighted that AI bots are not reliable for medical advice, but Anthropic and Google are doing something similar for health related features. Overall, Ashley, what is your take on this situation?
Ashley Coffey (07:23.865)
This is crazy. And you know, I see more lawsuits popping up every single day. And we've covered, especially the one on character.ai. But I also want to highlight here that just a day ago, Sam Altman went on a podcast and basically said, We are now like in the singularity. Which what is the AI singularity? Of course, there's no universally adopted edition, but it's essentially of he was talking about how.
It weighs on him that he knows that there's going to be a mass amount of displacement and people dying because of AI. And he essentially is like, I'm okay with that because I know the benefits outweigh the bad. And I'm like, no, we shouldn't be putting things out there that put people at risk. So that's really insightful and dangerous for a founder of an AI company to say those things.
And you can read online, it's a Forbes article. But also, yeah, these are extremely dangerous tools that we don't really have a lot of research on yet. And not to mention companies are hungry for data. I mean, they're, you know, scanning books and shredding them. What better data can they have than biometric data? It's gold for them. So no thank you. I'm not going to enable this. I would rather rely on the medical experts that went to school.
Daniel Hill (08:44.119)
I agree.
Ashley Coffey (08:52.182)
And, you know, know my chart and I've been in this field rather than something that's not human something that's not human should not be treating a human. Full stop. That's my that's my take.
Daniel Hill (09:02.579)
I understand. I don't a hundred percent agree. And I would say that my response is based on the fact that sometimes you want a second opinion, right? And I know that doctors and other medical professionals often use AI to catch things that they might have missed. And not that AI is infallible because it certainly is, and not that humans are infallible, because even medical professionals can make mistakes and miss things. But I think AI should be the backup, right? I like AI for.
explaining test results or defining medical terms after a doctor has given me that information, right? I sat with a doctor, I didn't catch everything that he said. Let me upload my test results and see what the AI says, how does it correspond, how does it line up with what the doctor said in terms of using it as a not second, but maybe third opinion. I like, you know, getting ideas of what a specific illness might be, but not for a diagnosis, right? I still want.
Ashley Coffey (09:45.754)
Mm.
Ashley Coffey (09:52.924)
Daniel Hill (09:59.476)
A doctor for a diagnosis, but as maybe a clue in the direction of what it could be, that's where I would leverage AI. That's my opinion. It's similar to how I feel about AI for mental health, right? Like I like it when you don't have any other option, but it's not a medical professional. It's not trained. It doesn't have the behavioral health training that you would need for a mental health therapist. So I view it similarly, you know, I view these two things kind of in the same light, but you know, certainly.
Ashley Coffey (10:07.77)
Yes.
Daniel Hill (10:27.698)
I recognize that I might be in the minority and thinking of it that way.
Ashley Coffey (10:31.694)
No, I I I agree with you in that respect because here's what I think of. I think of the time that I got automated notification that my test results from like an abdomen CT scan were available in my portal. And I could log in to see those results, but those results come to you automatically before your doctors had a chance to go over them with you. And if you've ever looked at like a CT scan or whatever, it's hard to interpret.
So I was thinking, like, you know, what stops someone from downloading that, putting it into a tool, and then using that at face value and getting freaked out because they're not having a conversation with a medical professional. Very small outlier case.
Daniel Hill (11:08.673)
Yes. And and no, but that's exactly something that someone would do. I myself have done that in the past. And then I went to the doctor and I listened to what they said with much more weight than what the AI told me. So I think it's perfectly, in my opinion, okay to do that if you feel comfortable to, as long as you're not giving it a hundred percent of the weight of what decision you're going to make based on the data that the AI is giving you, and you trust the human professional more in that scenario.
Ashley Coffey (11:26.129)
Yes.
Ashley Coffey (11:38.522)
Yes, one hundred percent.
Daniel Hill (11:40.738)
Yes. And I think it's okay. And again, this is my opinion, to bring what the AI tells you. And after you've heard what the doctor says, say, that makes sense. By the way, I ran it through AI and it said this because maybe the doctor missed that. That's definitely possible. Or maybe they will kick you out of the office and say, why are you coming here if you're just going to ask an AI? But I think that is good to have all the information available when you make a decision about what's right for your own medical health, provided you are trusting a professional.
Ashley Coffey (11:48.198)
Yeah.
Ashley Coffey (12:09.284)
Yeah. And a good professional will appreciate the fact that you one care about being involved in your health, but are willing to have a conversation to say, Hey, this is what I've gathered. Can you validate this or tell me if this is the right thing? 'Cause like that is a productive conversation versus get out of my office. so I yeah, I like that, yeah, for sure.
Daniel Hill (12:30.231)
That is it for our AI news you should know about this week. You can find links to everything we've discussed in the show notes. Please find us on social media and let us know what you think of the show. The links to connect with us on social media are in the show notes as well. You can find me at Daniel Hill Media on Instagram, YouTube, and Threads.
Ashley Coffey (12:47.812)
And if you want more tech talk, you can find me on Instagram at AshleyRcoffee89, where you can sign up for my free AI monthly newsletter that gives you prompts and different tips and tricks to level up your AI life.
Daniel Hill (13:01.719)
Thank you for listening and come back next week for more AI news you should know about.