What would turn bird flu - the kind that's killed millions of birds around the world and a few hundred people - into the next deadly pandemic?

Scientists want to know so they can get ahead of it. That's why, in 2010, two groups of researchers were studying an avian influenza virus that killed about half of the people it infected but does not spread easily among them. They infected ferrets with it to see what it would take to make it more transmissible. Ferrets' lungs and airways are a lot like ours.

'They would infect a ferret, wait a certain amount of time, take the virus that comes out of that ferret and infect the next ferret,' said University of Michigan microbiologist Michael Imperiale.

After 10 rounds, the virus had mutated into a much more dangerous form. 'They didn't have to transfer it from one ferret to the next,' he said. 'If they put the ferrets in close proximity, those ferrets could transfer to each other.'

That virus has pandemic potential.

The scientists figured out what parts of the virus had mutated to make it spread more easily. That could help experts looking out for the next dangerous virus, they said, and it could help develop drugs and vaccines against it.

FILE - A lab assistant in a laboratory of the Bavarian health office for animals prepares samples to be checked for the H5N1 virus, which causes the bird flu, Nov. 21, 2005 in Oberschleissheim near Munich, Germany.

On the other hand, the researchers had just created a new, deadly, easy-to-spread disease threat. If it were to get out of the lab somehow, it could start a global pandemic.

Lab leaks are rare but they do happen. SARS escaped from labs in Singapore, Taiwan and China in the early 2000s. A lab accident in Britain led to an outbreak of foot and mouth disease in livestock.

'It just shows that even the best labs in the world are not totally protected from accidents,' said Harvard University epidemiologist Marc Lipsitch. 'Humans are fallible.'

The bird flu experiments set off alarm bells when the researchers went to publish their results in 2011, according to Rutgers University microbiologist Richard Ebright.

'I think everyone hearing the result was surprised, first, that anyone would be so ethically challenged as to conduct such an experiment; and second, surprised that there were no mechanisms of oversight whatsoever that would prevent a researcher from conducting such an experiment,' he said.

Not everyone thought the experiments were fundamentally a bad idea. But they did kick off a push for more oversight.

That kind of research is worthwhile under two conditions, according to University of Michigan microbiologist Imperiale.

'Is it addressing a pressing biomedical issue, [for which] we have to have the answer? Is that really the only, or the best way to get it such that other approaches just will not get you to the answer you need? And if the answer to those two questions are yes, then let's figure out a way to do it safely,' he said.

U.S. health authorities set up a framework to ask those questions and decide what research should get funding.

However, Ebright said, 'that framework was sound on paper but it has not actually occurred in practice.'

He said that several research projects that should have gotten extra scrutiny did not, including U.S.-funded research in Wuhan, China on what would make coronaviruses more infectious to people. Ebright contends that research could have been the source of the COVID-19 virus.

Where COVID-19 came from is fiercely debated. But the coronavirus pandemic has drawn more scrutiny of research on germs with pandemic potential.

'It reminds us that this is no laughing matter. It does not have to be a 50 percent lethal virus to cause global societal disruption and a lot of damage and that we need to be really careful,' Lipsitch said.

'I think it has spurred some more cautious thought about oversight but the oversight really still is not in place,' he added.

A U.S. advisory board has recommended ways to tighten the guidelines. Experts say they are a step in the right direction but loopholes remain. Some say they should be law, not guidelines.

On the other hand, Lipsitch noted, 'there's a real risk that Congress will overreact and unduly restrict very valuable and completely safe research in an effort to restrict less valuable and more dangerous research.'

Plus, the rules would only apply to U.S.-funded research. Other countries need to do the same, Imperiale said.

'If this is not some sort of international consensus, we're not accomplishing what we really want to accomplish here, which is keeping the world safe,' he said. 'Because, as we all know, infectious agents don't respect national borders.'