OpenAI Faces Scrutiny Over Shortened AI Model Safety Testing Timelines

OpenAI partner says it had relatively little time to test the company's newest AI models | TechCrunch

An organization OpenAI frequently partners with to probe the capabilities of its models and evaluate them for safety, Metr, suggests that it wasn't given much time to test the company's powerful new releases, o3 and o4-mini. In a blog post published Wednesday, Metr writes that its red teaming of o3 was "conducted in a relatively short time" compared to the organization's benchmarking of a previous OpenAI flagship model, o1. This is significant, they say, because more testing time can lead to more comprehensive results. "This evaluation was conducted in a relatively short time, and we only tested the model with simple agent scaffolds," wrote Metr in a blog post. "We expect higher performance [on benchmarks] is possible with more elicitation effort." Recent reports suggest that OpenAI, spurred by competitive pressure, is rushing independent evaluations. According to the Financial Times, OpenAI gave some testers less than a week for safety checks for an upcoming major release. In statements, OpenAI has disputed the notion that it's compromising on safety. Metr says that, based on the information it was able to glean in the time it had, o3 has a "high propensity" to "cheat" or "hack" tests in sophisticated ways in order to maximize its score -- even when the model clearly understands its behavior is misaligned with the user's (and OpenAI's) intentions. The organization thinks it's possible o3 will engage in other types of adversarial or "malign" behavior, as well -- regardless of the model's claims to be aligned, "safe by design," or not have any intentions of its own. "While we don't think this is especially likely, it seems important to note that this evaluation setup would not catch this type of risk," Metr wrote in its post. "In general, we believe that pre-deployment capability testing is not a sufficient risk management strategy by itself, and we are currently prototyping additional forms of evaluations." Another of OpenAI's third-party evaluation partners, Apollo Research, also observed deceptive behavior from o3 and o4-mini. In one test, the models, given 100 computing credits for an AI training run and told not to modify the quota, increased the limit to 500 credits -- and lied about it. In another test, asked to promise not to use a specific tool, the models used the tool anyway when it proved helpful in completing a task. In its own safety report for o3 and o4-mini, OpenAI acknowledged that the models may cause "smaller real-world harms" without the proper monitoring protocols in place. "While relatively harmless, it is important for everyday users to be aware of these discrepancies between the models' statements and actions," wrote the company. "[For example, the model may mislead] about [a] mistake resulting in faulty code. This may be further assessed through assessing internal reasoning traces."

OpenAI used to test its AI models for months - now it's days. Why that matters

On Thursday, the Financial Times reported that OpenAI has dramatically minimized its safety testing timeline. Also: The top 20 AI tools of 2025 - and the No. 1 thing to remember when you use them Eight people who are either staff at the company or third-party testers told FT that they had "just days" to complete evaluations on new models -- a process they say they would normally be given "several months" for. Evaluations are what can surface model risks and other harms, such as whether a user could jailbreak a model to provide instructions for creating a bioweapon. For comparison, sources told FT that OpenAI gave them six months to review GPT-4 before it was released -- and that they only found concerning capabilities after two months. Also: Is OpenAI doomed? Open-source models may crush it, warns expert Sources added that OpenAI's tests are not as thorough as they used to be and lack the necessary time and resources to properly catch and mitigate risks. "We had more thorough safety testing when [the technology] was less important," one person, who is currently testing o3, the full version of o3-mini, told FT. They also described the shift as "reckless" and "a recipe for disaster." Also: This new AI benchmark measures how much models lie The sources attributed the rush to OpenAI's desire to maintain a competitive edge, especially as open-weight models from competitors, like Chinese AI startup DeepSeek, gain more ground. OpenAI is rumored to be releasing o3 next week, which FT's sources say rushed the timeline to under a week. The shift emphasizes the fact that there is still no government regulation for AI models, including any requirements to disclose model harms. Companies including OpenAI signed voluntary agreements with the Biden administration to conduct routine testing with the US AI Safety Institute, but records of those agreements have quietly fallen away as the Trump administration has reversed or dismantled all Biden-era AI infrastructure. Also: OpenAI research suggests heavy ChatGPT use might make you feel lonelier However, during the open comment period for the Trump administration's forthcoming AI Action Plan, OpenAI advocated for a similar arrangement to avoid navigating patchwork state-by-state legislation. Also: The head of US AI safety has stepped down. What now? Outside the US, the EU AI Act will require that companies risk test their models and document results. "We have a good balance of how fast we move and how thorough we are," Johannes Heidecke, head of safety systems at OpenAI, told FT. Testers themselves seemed alarmed, though, especially considering other holes in the process, including evaluating the less-advanced versions of the models that are then released to the public or referencing an earlier model's capabilities rather than testing the new one itself. Also: The Turing Test has a problem - and OpenAI's GPT-4.5 just exposed it Get the morning's top stories in your inbox each day with our Tech Today newsletter.

OpenAI slashes AI model safety testing time

OpenAI has slashed the time and resources it spends on testing the safety of its powerful artificial intelligence models, raising concerns that its technology is being rushed out without sufficient safeguards. Staff and third-party groups have recently been given just days to conduct "evaluations", the term given to tests for assessing models' risks and performance, on OpenAI's latest large language models, compared to several months previously. According to eight people familiar with OpenAI's testing processes, the start-up's tests have become less thorough, with insufficient time and resources dedicated to identifying and mitigating risks, as the $300bn start-up comes under pressure to release new models quickly and retain its competitive edge. "We had more thorough safety testing when [the technology] was less important," said one person currently testing OpenAI's upcoming o3 model, designed for complex tasks such as problem-solving and reasoning. They added that as LLMs become more capable, the "potential weaponisation" of the technology is increased. "But because there is more demand for it, they want it out faster. I hope it is not a catastrophic mis-step, but it is reckless. This is a recipe for disaster." The time crunch has been driven by "competitive pressures", according to people familiar with the matter, as OpenAI races against Big Tech groups such as Meta and Google and start-ups including Elon Musk's xAI to cash in on the cutting-edge technology. There is no global standard for AI safety testing, but from later this year, the EU's AI Act will compel companies to conduct safety tests on their most powerful models. Previously, AI groups, including OpenAI, have signed voluntary commitments with governments in the UK and US to allow researchers at AI safety institutes to test models. OpenAI has been pushing to release its new model o3 as early as next week, giving less than a week to some testers for their safety checks, according to people familiar with the matter. This release date could be subject to change. Previously, OpenAI allowed several months for safety tests. For GPT-4, which was launched in 2023, testers had six months to conduct evaluations before it was released, according to people familiar with the matter. One person who had tested GPT-4 said some dangerous capabilities were only discovered two months into testing. "They are just not prioritising public safety at all," they said of OpenAI's current approach. "There's no regulation saying [companies] have to keep the public informed about all the scary capabilities . . . and also they're under lots of pressure to race each other so they're not going to stop making them more capable," said Daniel Kokotajlo, a former OpenAI researcher who now leads the non-profit group AI Futures Project. OpenAI has previously committed to building customised versions of its models to assess for potential misuse, such as whether its technology could help make a biological virus more transmissible. The approach involves considerable resources, such as assembling data sets of specialised information like virology and feeding it to the model to train it in a technique called fine-tuning. But OpenAI has only done this in a limited way, opting to fine-tune an older, less capable model instead of its more powerful and advanced ones. The start-up's safety and performance report on o3-mini, its smaller model released in January, references how its earlier model GPT-4o was able to perform a certain biological task only when fine-tuned. However, OpenAI has never reported how its newer models, like o1 and o3-mini, would also score if fine-tuned. "It is great OpenAI set such a high bar by committing to testing customised versions of their models. But if it is not following through on this commitment, the public deserves to know," said Steven Adler, a former OpenAI safety researcher, who has written a blog about this topic. "Not doing such tests could mean OpenAI and the other AI companies are underestimating the worst risks of their models," he added. People familiar with such tests said they bore hefty costs, such as hiring external experts, creating specific data sets, as well as using internal engineers and computing power. OpenAI said it had made efficiencies in its evaluation processes, including automated tests, which have led to a reduction in timeframes. It added there was no agreed recipe for approaches such as fine-tuning, but it was confident that its methods were the best it could do and were made transparent in its reports. It added that models, especially for catastrophic risks, were thoroughly tested and mitigated for safety. "We have a good balance of how fast we move and how thorough we are," said Johannes Heidecke, head of safety systems. Another concern raised was that safety tests are often not conducted on the final models released to the public. Instead, they are performed on earlier so-called checkpoints that are later updated to improve performance and capabilities, with "near-final" versions referenced in OpenAI's system safety reports. "It is bad practice to release a model which is different from the one you evaluated," said a former OpenAI technical staff member. OpenAI said the checkpoints were "basically identical" to what was launched in the end.

OpenAI cuts back on AI model safety testing- FT By Investing.com

Investing.com-- OpenAI has slashed the amount of time and resources spent on testing the safety of its artificial intelligence models, raising some concerns over proper guardrails for its technology, the Financial Times reported on Friday. Staff and groups that assess risks and performance of OpenAI's models were recently given just days to conduct evaluations, the FT report said, citing eight people familiar with the matter. The start-up's testing processes have become less thorough with fewer resources and less time dedicated towards mitigating risks, the FT report said. The report comes as OpenAI races to release updated AI models and maintain its competitive edge, especially amid heightened competition from new Chinese entrants such as DeepSeek. OpenAI is gearing up to release its new o3 model by next week, although no release data has been determined. But this rush to release updated models is potentially compromising the firm's safety checks. Still, reports of lower safety testing times also come amid a shift in AI towards inference- ie the processing and generation of novel data- from training, which uses existing data to improve the capabilities of an AI model. OpenAI said earlier in April that it had raised $40 billion in a funding round led by Japan's SoftBank Group Corp. (TYO:9984), which valued the company at $300 billion.