Adm. Bobby R. Inman, deputy director of the CIA, warned scientists yesterday that they face a government crackdown to curb Soviet use of militarily sensitive American technology unless they agree to voluntary "reviews" of their work by intelligence agencies.

If scientists do not cooperate in keeping some of their papers secret voluntarily, they will encounter a "tidal wave" of public outrage resulting in tough restrictive laws, Inman told a panel at the annual meeting of the American Association for the Advancement of Science.

Scientists should beware that congressional investigations now in progress will point up the "thoroughly documented" fact that, in the buildup of Soviet defense capability, "the bulk of new technology which they have employed has been acquired from the United States," Inman said.

When the details of this "hemorrhage of the country's technology" become known, Inman said, public outrage will lead to laws restricting the publication of scientific work that the government might consider "sensitive" on national security grounds.

Most of the audience consisted of military officers and businessmen who appeared to sympathize with Inman's proposal. He got hostile questions, however, from the handful of scientists present. They considered the proposal repressive censorship.

"The tides are moving, and moving fast, toward legislated solutions that in fact are likely to be much more restrictive, not less restrictive, than the voluntary" censorship system he has suggested, Inman said.

When he was director of the National Security Agency, the codemaking and breaking intelligence agency, Inman led an effort to have prominent private researchers submit their papers on the mathematical theory of codes before publication.

The NSA also briefly imposed secrecy orders on some private code research in recent years.

But in April, 1981, the National Science Foundation, the American Council on Education and the NSA cooperatively produced a voluntary review system under which scientists can submit their papers to the NSA and receive a judgment on whether they possibly contain information damaging to the national security.

Since then, about 25 papers have been reviewed and none had problems, according to Daniel Schwartz, until recently chief counsel for the NSA.

Inman wants to extend this sort of voluntary system to many other kinds of work, he said yesterday.

"There are other fields where publication of certain information could affect the national security in a harmful way," Inman said. He cited "computer hardware and software, other electronic gear and techniques, lasers, crop projections, and manufacturing procedures."

Rather than a faceoff between scientists and the protectors of national security, he said, "I believe a wiser course is possible....A potential balance between national security and science may lie in an agreement to include in the peer review process, prior to the start of research and prior to publication, the question of potential harm to the nation."

He did not go into detail except to say that he would like to modify in some way the manner in which scientific work and papers are normally reviewed to allow intelligence agencies access to the system.

Inman said one problem in getting cooperation from scientists is that intelligence agencies usually cannot explain why they want to censor a particular publication, or even define the kind of information they want to censor, because this may be as revealing as the publication itself.

But he warned that those who say "don't give us any regulations" are "about to have that way of thinking washed away by the tidal wave" of public outrage.