What if we’re wrong on Iran?

As the U.S. contemplates whether to attack Iran’s nuclear facilities, intelligence community leaders should be asking themselves a question: What if we’re wrong?

That question wasn’t asked — or at least wasn’t answered — in the run-up to the 2003 Iraq war, with devastating consequences. Before giving up on containment or deterrence polices and undertaking a “preventive war” against a nation that has not attacked the United States, we should be as certain as possible of the evidence.

Iran today presents an even murkier intelligence picture than Iraq did in 2003. We have not had a diplomatic presence there since 1979 and have had to rely on intelligence collected through technology, international inspectors and foreign intelligence relationships. In the absence of solid intelligence, the intelligence community has had to fall back on its own assumptions or mind-sets regarding Iran’s nuclear program and make educated guesses about how its government would probably operate its programs. Our assessments of Iran’s military capabilities have had to rely in part on our understanding of how American weapon developers conduct tests and develop weapons. Such analytic assumptions have and can again lead to incorrect conclusions.

So what can be done to avert another military strike based on devastatingly wrong intelligence estimates? How can we avoid terrible mistakes, followed by commission investigations and finger-pointing?

First, we must set extremely high standards for evidence. Intelligence professionals must challenge themselves to look at their forecasts skeptically, asking whether underlying assumptions about Iranian behavior and technical talent are well founded. Policymakers should encourage this kind of critical thinking and be attentive to signs of faulty logic or flawed intelligence.

Second, the U.S. must not over-rely on information gathered and supplied by foreign governments. Such intelligence can be useful, but it is often provided as much to influence action by American policymakers as to provide unbiased and accurate information. Much of the foreign intelligence used to analyze Iraq’s supposed development of weapons of mass destruction, for example, was self-serving, biased or flat-out fabricated. The U.S. is almost certainly the target for influence operations designed to shape American perceptions of the Iranian nuclear program. We should be especially wary of reporting that fits what we are expecting to see and challenge that reporting vigorously.

Third, U.S. intelligence personnel should be kept at arm’s length from policy discussions, particularly those involving military options. Being too close to the decisions made by the Bush administration more than likely contributed to then-CIA Director George Tenet’s assertion that the U.S. had “slam-dunk” evidence that Saddam Hussein was actively developing weapons of mass destruction. Senior intelligence officials are not immune from wanting to be team players, and if a policy team is looking for information to support a desired action, this can skew an intelligence agency’s views on the information it has gathered. James R. Clapper, the U.S. director of national intelligence, would be well advised to run all intelligence about Iran’s nuclear weapons program through a rigorous “red-teaming” exercise, involving outside experts who have no ax to grind or connections to the current administration or its policies.

Fourth, the intelligence community should never be called on to make the case for intervention, as was the situation in 2003, when so-called white papers on Iraq’s WMD program were fashioned by the intelligence community to support Bush administration policies. Such reports are not rigorous intelligence assessments but rather advocacy pieces devoid of the important qualifiers that coordinated intelligence reports should carry.

Finally, the intelligence community should immediately, if it has not already done so, prepare candid assessments of the effect military strikes against Iran’s nuclear facilities could have on both Iranian politics and regional stability. In 2003, such assessments proved to be prescient, but they were completed only after the decision had been made to invade Iraq. Consequently, they had virtually no impact on decision-makers, who had convinced themselves that Iraqis would greet us as liberators and quickly restore the functioning of their society and economy.

Needless to say, the decision to attack another Islamic state would carry consequences far beyond reducing Iran’s military potential, and the intelligence community needs to analyze those consequences concurrently with its analysis of intelligence regarding Iran’s nuclear intentions.

These steps will not guarantee that intelligence used to reach the important decisions regarding Iran will be perfect. Clearly, it will not be. However, the intelligence community should not repeat mistakes it made in 2002 and 2003, nor allow itself to become the scapegoat for decisions that properly reside with the nation’s political and military leadership.

Roger Z. George, a former national intelligence officer, teaches at the National War College in Washington.