WASHINGTON, March 1 — For more than three years, American intelligence officials have insisted that they learned from their mistakes in the months leading to the Iraq war, when murky information about Saddam Hussein’s weapons programs was presented as fact and inconclusive judgments were hardened into statements of near certainty.
The more calibrated intelligence assessments that have come to light in recent weeks, particularly on Iran and North Korea, appear to show a new willingness by American spy agencies to concede the limits of their knowledge.
The new caution reflects adherence to what some officials now call “the Powell Rule.” That rule is intended to avoid a repetition of former Secretary of State Colin L. Powell’s humiliation after the satellite photos and intercepted communications he presented to the United Nations Security Council as proof that Iraq was stockpiling banned weapons turned out to be nothing of the sort.
It was in large part because of the Iraq intelligence debacle that top intelligence officials over the past year have overhauled the way in which they pass judgments about some of the world’s most difficult spying targets, including Iran and North Korea. The new operating approach is intended to ensure that intelligence is solid before it is publicly presented, and even then, that clear distinctions are made between facts and inferences.
Intelligence professionals have several motivations for the new approach. Those include protecting their own credibility and ensuring that assessments can stand up to careful scrutiny in the wake of multiple government investigations condemning the prewar intelligence assessments about Iraq. Officials also said that greater caution had become ever more necessary because of increasing inexperience in the work force; about 50 percent of current intelligence analysts have less than five years of experience.
At the same time, the use of specific, careful language is seen by intelligence professionals as a barrier that can help to prevent intelligence assessments from being exaggerated, twisted or otherwise misused by policy makers to advance specific agendas.
Under the new guidelines, the assessments produced by intelligence agencies must include detailed descriptions of the subjects on which analysts disagree, and analysts must now “show their work” by including in their judgments the chain of logic that led them to their conclusions.
Intelligence analysis has always been more art than science, and in recent statements, American intelligence officials have laid bare the limits of what they believe they know about Iran and North Korea. Officials have said they are certain that Iranian paramilitary groups are providing lethal bomb material to Iraqi Shiite groups, but they do not know whether this support has been approved at the highest levels of government in Tehran.
In testimony before the Senate Armed Services Committee this week, Mike McConnell, the new director of national intelligence, said that it was possible Iran could produce a nuclear weapon by the beginning of the next decade, but that it was also possible Iran would not develop a weapon until 2015. At the same hearing, another top official acknowledged publicly that the American intelligence agencies were less certain today than they were five years ago about whether North Korea was pursuing a clandestine uranium enrichment program.
John D. Negroponte formalized the new analytical procedures in a directive issued in January, one of his last acts as director of national intelligence. The directive stated that “the analytic process must be as transparent as possible” and that “analysis must be objective and independent of political considerations.”
The pre-emptive war doctrine that President Bush has articulated since the Sept. 11 attacks holds that the United States has an obligation take action against security threats before they fully materialize. A corollary embraced by the White House has held that policy makers must assume the worst about the intentions of adversaries, even with imperfect intelligence about their intentions and capabilities.
That worst-case approach was very much on display when Mr. Bush said with certainty in 2002 and 2003 that Iraq possessed chemical and biological weapons. On North Korea, Mr. Bush said in November 2002 that “contrary to an agreement they had with the United States, they’re enriching uranium, with the desire of developing a weapon.”
As intelligence agencies have backed away from those views, some former officials worry that the Iraq experience may cause the pendulum to swing too far in the direction of circumspection. They say that they hope the administration does not accept at face value the denials from North Korea and Iran about their banned weapons programs.
“I worry about that, but I am not willing to accept that doomsday is upon us,” said John R. Bolton, who stepped down in December as ambassador to the United Nations after it was clear he would not win Senate confirmation. Mr. Bolton said he still believed that policy makers would not allow intelligence analysts’ new caution to soften the United States’ policies toward its enemies too much.
The intelligence community’s renewed emphasis on debate and dissent has been reflected in decisions about which officials earned promotions in the aftermath of the Iraq intelligence failures.
Thomas Fingar, who once ran the State Department’s Bureau of Intelligence and Research, is now in charge of all analysis for the director of national intelligence.
It was Mr. Fingar’s State Department office in 2002 that produced one of the most significant dissents to that year’s National Intelligence Estimate, now discredited, challenging the majority view that Iraq had bought aluminum tubes to build up its nuclear weapons program.
During an interview late last year, Mr. Fingar said it was now his job to ensure there was as much competitive analysis as possible before intelligence reports were completed, if only to avoid a repeat of the 2002 N.I.E. findings, which were pivotal to the Senate vote that authorized the Iraq war.
“The basic idea is to avoid a premature rush to an artificial consensus,” Mr. Fingar said. “The interesting thing is not when analysts agree. It’s when they disagree.”