ChatGPT will now combat bias with new measures put forth by OpenAI
The artificial intelligence (AI) company recently unveiled an updated Model Spec, a document that defines how OpenAI wants its models to behave in ChatGPT and the OpenAI API. The company says this iteration of the Model Spec builds on the foundational version released last May.
"I think with a tool as powerful as this, one where people can access all sorts of different information, if you really believe we're moving to artificial general intelligence (AGI) one day, you have to be willing to share how you're steering the model," Laurentia Romaniuk, who works on model behavior at OpenAI, told Fox News Digital.
"People should be allowed to know what goes into the way these models respond and how the thoughts coming out of the model are crafted," she continued.
While some have argued that GPT-4o, the latest version of the technology, appears close to AGI, others say it will be years or decades before the technology reaches human-like abilities.
There is no single agreed upon definition of AGI, but a 2020 report from consulting giant McKinsey said a true AGI would need to master skills like sensory perception, fine motor skills, and natural language understanding.
Today, generative AI is the dominant form of this technology, with the ability to produce content including text, images and more. Generative AI, like chatbots, uses datasets with specific information to complete tasks and cannot go beyond the provided data.
They are also susceptible to bias in their datasets, whether intentional or accidental.
To better understand real-world performance and address bias, OpenAI has begun measuring progress by gathering a challenging set of prompts designed to test how well the models adhere to each principle in the Model Spec, essentially acting as a testing metric.
According to Joanne Jang, who leads the product for model behavior at OpenAI, the challenge with large language models is that they are rarely deterministic.
Jang stressed that one benefit of the Model Spec is that it clarifies the intended behavior, so everyone can understand and then debate it.
"When there are bugs in the behavior where a model output doesn't resonate or doesn't align with the [Model] Spec, then the public knows this is something we're working towards, and it's an ongoing area of science," she told Fox News Digital.
OpenAI says they also attempt to assume an objective point of view in their AI prompts and consciously avoid any agenda. For example, when a user asks if it is better to adopt a dog or get one from a breeder, ChatGPT provides both sides of the argument, highlighting the pros and cons of each.
According to OpenAI, a non-compliant AI answer that violates the Model Spec would provide what it believes to be the better choice and engage in an "overly moralistic tone" that might alienate those considering breeders for valid reasons.
As each AI system advances, OpenAI says it will iterate these principles, invite community feedback, and share progress openly.
OpenAI released this version of the Model Spec into the public domain under a Creative Commons license to support broad use and collaboration. This means developers and researchers can freely use and adapt the current metrics and help improve model behavior.
OpenAI says this update reinforces its belief in open exploration and discussion, with an emphasis on user and developer control and guardrails to prevent harm.
Romaniuk concludes that public discourse cannot exist without transparency, reinforcing the need for the Model Spec and community engagement.
"Ultimately, we believe in the intellectual freedom to think, speak and share without restriction. We want to make sure that users have that ability and that's what it's all about," she said.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
New York Post
4 minutes ago
- New York Post
America's wealthiest neighborhood is probably not where you'd expect it to be
Move over, Beverly Hills – there's a new neighborhood taking the crown as America's most expensive. Gables Estates, a small gated community located in Coral Gables, Florida, has now taken Zillow's top spot for the United States' most expensive neighborhood, based on home value data over the last 12 months. Advertisement The neighborhood now tops Beverly Hills, California, long regarded as the pinnacle ZIP code of wealth. 'What makes Gables Estates unique is its privacy, sophistication, and livability, which Beverly Hills doesn't fully offer,' Coral Gables-based CEO and co-founder of Vertical Developments, Fernando de Nunez y Lugones, told Fox News Digital. 'Unlike Beverly Hills, which is known mostly for its name and history, Gables Estates is about the lifestyle and the experience, and that consistently draws ultra-high-net-worth buyers.' Seven of the 10 priciest neighborhoods are located in the Sunshine State, while the remaining three are in California. 'This is a long-term trend, not a blip,' Lugones said. 'I remember reading last year that Coral Gables actually beat Beverly Hills, and it makes sense. Buyers and businesses are relocating here for lifestyle, taxes and the overall environment. South Florida is increasingly becoming known as Wall Street South, and Coral Gables stands out for its walkability, amenities, and culture. Markets may fluctuate, but the appeal of space, security, and community remains strong.' Advertisement Zillow's home value index is 'designed to capture the value of a typical property across the nation or the neighborhood,' its website states, by using metrics such as sales transactions, tax assessments, public records, square footage, and location. 5 Gables Estates, a small gated community located in Coral Gables, Florida, has now taken Zillow's top spot for the United States' most expensive neighborhood, according to reports. LightRocket via Getty Images The greater Miami area has experienced an influx of wealth migrating in the post-COVID era. Between 2017 and 2022, more than $14 billion in income flocked to Florida, with more than $9.2 billion going to Palm Beach, Broward, and Miami-Dade counties. Advertisement Additionally, Henley & Partners World's Wealthiest Cities Report for 2025 found that both West Palm Beach and Miami surpassed New York City as the world's fastest-growing wealth hubs. West Palm saw a 112% increase in millionaire growth over the last decade, while Miami saw a 94% increase. Gables Estates is an exclusive waterfront community that features luxurious mansions, lush landscaping, and a canal system that connects directly to Biscayne Bay. The neighborhood began development in the 1920s, and now includes an estimated 160 to 180 properties. Advertisement 5 'What makes Gables Estates unique is its privacy, sophistication, and livability, which Beverly Hills doesn't fully offer,' Coral Gables-based CEO and co-founder of Vertical Developments, Fernando de Nunez y Lugones, said. Bloomberg via Getty Images 'We expect even more interest from ultra-luxury buyers looking to move to Coral Gables and the broader Miami area. As more buyers settle here, the neighborhood's reputation as a world-class enclave will only continue to grow,' Lugones noted. Its residents enjoy high-level security with 24/7 armed guards on land and water. Entry into the community means paying a $100,000 non-refundable application fee to the larger Gables Estates Club. Zillow notes that home listing prices often exceed $21 million. 'Coral Gables is naturally self-limiting since there's only so much prime land, which helps protect against overexposure,' Lugones pointed out. 'At the same time, its prestige and careful development standards allow it to handle national attention without compromising lifestyle or value. As long as growth remains measured, Coral Gables can sustain its momentum for decades.' 5 Gables Estates is an exclusive waterfront community that features luxurious mansions, lush landscaping, and a canal system that connects directly to Biscayne Bay, according to reports. Universal Images Group via Getty Images 5 Entry into the community means paying a $100,000 non-refundable application fee to the larger Gables Estates Club. Carlos Barrios America's top 10 most expensive and least expensive neighborhoods can be found below: Most expensive neighborhoods in the U.S.: Advertisement 1. Gables Estates, Coral Gables, FL 2. Port Royal, Naples, FL 3. Old Cutler Bay, Coral Gables, FL 4. Beverly Hills Gateway, Beverly Hills, CA Advertisement 5. The Flats, Beverly Hills, CA 5 'We expect even more interest from ultra-luxury buyers looking to move to Coral Gables and the broader Miami area. As more buyers settle here, the neighborhood's reputation as a world-class enclave will only continue to grow,' Lugones noted. REUTERS 6. Shady Canyon, Irvine, CA 7. San Marino Island, Miami Beach, FL Advertisement 8. Bear's Club, Jupiter, FL 9. Palm Island, Miami Beach, FL 10. Rivo Alto Island, Miami Beach, FL Advertisement On Zillow's list, the least expensive neighborhood is Bullard Hill in Jackson, Mississippi, with an average home value of $24,026. Other neighborhoods include two in Shreveport, Louisiana; three in Flint, Michigan; and three in Jackson, Mississippi.
Business Insider
5 minutes ago
- Business Insider
Short-seller Andrew Left says OpenAI's $500 billion valuation proves Palantir stock is way too expensive
A week after Andrew Left criticized Palantir Technologies as highly overvalued, the short-seller is back with a more detailed analysis. On August 18, Left's investment firm, Citron Research, published a note on Palantir stock, holding it up against another AI titan to illustrate why its high valuation should be alarming to investors. In a Fox Business appearance last week, Left discussed his Palantir short position, stating that the stock wouldn't become cheap until it reached a price range of $40-$50 per share. However, since then, his firm has updated its view. Citron Research said in the note that even if Palantir fell to $40 a share—which would mark a 77% decline from Tuesday's high—it would still be expensive. The firm highlighted OpenAI's plan to sell $6 billion worth of stock to illustrate its view on Palantir. "Its new $500 billion valuation provides a true benchmark for evaluating Wall Street's favorite trading stock, Palantir, a company now detached from fundamentals and analysis, ironically the very services it claims to offer," Citron stated. The firm provided a breakdown of why Palantir would be lucky to reach OpenAI's valuation multiple. It began by dividing Palantir's projected 2026 revenue of roughly $500 billion by OpenAI's roughly $29.6 billion, resulting in a price-to-revenue multiple of 17x. "If Palantir trades at the same 17x price-to-revenue multiple on $5.6 billion in revenue (with approximately 2.37 billion shares outstanding), the implied stock price would be about $40 per share." However, the note went on to say that even if Palantir stock were to fall to that level, it would still be one of the most expensive software as a service (SaaS) stocks in history. Left questioned whether Palantir deserves to be trading at the same multiple as OpenAI, making it clear that he does not think it belongs in the same league as the AI titan. "What's remarkable about OpenAI isn't just its rapid growth, but the unprecedented scale at which it's occurring—unique in tech history," the note said. After reporting strong second-quarter earnings and rising more than 118% year-to-date, Palantir stock has declined over the past week. Shares continued to slide on Tuesday, down about 6% to $163.92 at midday.
WIRED
5 minutes ago
- WIRED
OpenAI Is Poised To Become The Most Valuable Startup Ever. Should It Be?
Aug 19, 2025 12:00 PM The AI startup is chasing a $500 billion valuation, with backers betting it can become the next Apple or Google. There are reasons for skepticism. Photo-illustration: WIRED Staff; Getty Images OpenAI is reportedly on the verge of a roughly $500 billion valuation, a figure that would make it the most valuable private company in the world—bigger than SpaceX, TikTok's parent company Bytedance, and even public giants like Palantir. It's a staggering number for a company with an 'astronomical burn rate.' How is this even possible? As Axios reports, there are actually two deals in play: a SoftBank-led round valuing the company at $300 billion, which won't close until year's end, and a secondary sale of employee shares at a far steeper $500 billion valuation. Most of the cheaper shares have already been snapped up, leaving investors to fight over the pricier ones. One OpenAI investor—who spoke on the condition of anonymity, citing an NDA—compared it to the dawn of the internet. 'We're in one of the biggest technology shifts [in history],' the investor tells me. 'The outcomes continue to get bigger than people think.' The investor argues the math for investing at the $500 billion valuation is straightforward: Hypothetically, if ChatGPT hits two billion users and monetizes at $5 per user per month—"half the rate of things like Google or Facebook'—that's $120 billion in annual revenue. 'That alone would support a trillion and a half dollar company, which is a pretty good return, just thinking about ChatGPT,' the investor says. 'it doesn't include all the rest of stuff they're working on, all the enterprise stuff, all the agentic stuff, all of the work they're doing on hardware.' Trillions of Dollars The $5 figure is, admittedly, back of the envelope math. Today, ChatGPT has 700 million weekly active users—and fewer than 10 percent pay (OpenAI declined to comment on this figure). The investor's projections are ambitious, and seem to discount the threat of major players like Google and Meta eating OpenAI's lunch. 'The half a trillion dollar question now is, to what extent will OpenAI be able to retain the customers that it has acquired, and simultaneously be able to bring its costs to a point where it can, in fact, monetize at [hypothetically] $5 per user per month,' says Arun Sundararajan, a professor at New York University's Stern School of Business. The bet here is that OpenAI is the next Facebook or Google. For investors buying in at $500 billion, 'they're expecting an IPO above a trillion in two to three years, otherwise the rate of return does not justify the investment,' says Glenn Okun, who's also a business professor at NYU. That would mean leaping into the top ten most valuable public companies in the world almost overnight. The investor says they have a longer time horizon than that, but 'of course an IPO is the most sensible path given the scale of the company.' Though the investor admits, yes, the company would need to be valued at more than $1 trillion to make the investment worthwhile. Stranger things have happened—particularly to OpenAI. In the first seven months of 2025, the company doubled its projected annual revenue to $12 billion, which would estimate OpenAI is bringing in about $1 billion per month. Enterprise adoption has surged too, reaching 5 million paying business users this month. Not to mention what potential advertising revenue could do to its bottom line. To the investor, these are signs of a company with the momentum to win: 'People don't like unprecedented things, because most people like to pattern match,' the investor says. 'Everything this company has done has been unprecedented from the pace of its revenue growth to the AI technology.' It's a nice dream. But OpenAI's spending is ballooning alongside its revenue. It's reportedly expecting a cash burn of $8 billion this year, and Altman said at a recent dinner in San Francisco that much of OpenAI's infrastructure costs are for inference—and he expects the startup to spend trillions of dollars on datacenters in the 'not very distant future.' That means in the fantastical future where ChatGPT may have billions of users, it would cost trillions of dollars to serve them. It's extraordinarily expensive to scale such a power hungry technology. Even if the chips get faster and cheaper, the overall costs grow as the company works to train even bigger models and serve them to a growing user base sending off billions of queries, Sundararajan tells me. The investor, however, brushes off profitability concerns, saying OpenAI can spread those massive costs across its entire user base. 'It is not cheap, and it's not for the faint of heart in terms of investment,' they add. Okun takes a more skeptical approach. 'What we're looking at here is major capital investment with the sorts of valuations that may be very hard to to actually reach,' he says. A New Bubble The $500 billion headline also has caveats. Private-market pricing reflects what a small number of investors are willing to pay for a small number of shares. In other words, the feeding frenzy is factored into the share price. 'In all likelihood, it's an approximation,' Okun says of the valuation. He adds that the number reflects both firm value and the competition between investors to get a slice of the pie. 'We really can't accurately disentangle those two streams of value in order to isolate firm value.' Right now, investors appear to have a nearly insatiable appetite for AI startups. According to recent PitchBook data, a third of all venture dollars in the second quarter of this year went to five companies—all in the AI space—which is more than double the spending in 2024. Is that a bubble? Altman admitted as much at the dinner. But in classic Altman fashion, he reminded attendees that's actually a good thing: 'When bubbles happen, smart people get overexcited about a kernel of truth,' he said. 'If you look at most of the bubbles in history [like] the tech bubble, there was a real thing. Tech was really important, the internet was a really big deal.' Startup valuations have always reflected how well the CEO can sell a vision. At OpenAI, Altman cranks it up to eleven, and it's seemingly paying off. Meanwhile, investors are betting that ChatGPT becomes as indispensable as Google, that billions of people happily pay to use it, and competitors quietly fade into irrelevance. Whether any of that actually happens almost doesn't matter. For now, the story is worth half a trillion all on its own. 'There's definitely risk in it,' the investor tells me. 'The story is not fully written.' This is an edition of Kylie Robison's Model Behavior newsletter . Read previous newsletters here.
