Al-Qaida has been seeking to acquire nuclear weapons since the early 1990s, when Osama Bin Laden got scammed to the tune of $1.5 million while trying to buy weapons-grade uranium. One month before 9/11, Bin Laden and al-Zawahiri met with Sultan Bashiruddin Mahmood, a key architect of Pakistan's nuclear program, known to possess nutty end-of-days beliefs about nukes and Islam. (Mahmood insisted, implausibly, that his sole purpose was to solicit funds for a polytechnic university he wished to establish in Kabul but that Bin Laden kept pestering him about building al-Qaida a nuclear bomb, which he refused to do. He's been under Pakistani surveillance ever since.) At the meeting, Bin Laden told Mahmood that he'd acquired nuclear material from Uzbekistan but that it wasn't of sufficient grade to make a nuclear weapon. News of this meeting helped fuel a flurry of panic at the Central Intelligence Agency in October 2001 over a report, later proved untrue, that al-Qaida had acquired a 10-megaton bomb stolen from Russia's nuclear arsenal. Al-Zawahiri has since boasted that al-Qaida possesses nuclear weapons, but that's highly doubtful. Pakistan's recent release from house arrest of nuclear scientist A.Q. Khan, who sold nuclear secrets to North Korea, Iran, and Libya, probably increased the risk of nuclear proliferation, but it's hard to know by how much.
Graham Allison, a Harvard political scientist of some renown, wrote in his 2004 book Nuclear Terrorism that "a nuclear terrorist attack on America in the decade ahead is more likely than not." When the paperback came out, he wrote in an afterword that "the likelihood, indeed inevitability, of a nuclear terrorist attack absent a major departure for current policy and practice" had increased over the previous year. In "World At Risk," a report about proliferation and terrorism released in December 2008, Allison and his fellow members of a congressional blue-ribbon panel pushed the deadline back to 2013, broadened the location to "somewhere in the world," and broadened the weapons category to include biological and chemical agents. Such predictions cause other terrorism experts to roll their eyes. John Mueller, a political scientist at Ohio State who believes the terrorism threat is overstated, twitted Allison for predicting as far back as 1995 that "acts of nuclear terrorism against American targets before this decade [i.e., the 1990s] is out."
In fact, the likelihood of nuclear terrorism isn't that great. Mueller points out that Russian "suitcase bombs," which figure prominently in discussions about "loose nukes," were all built before 1991 and ceased being operable after three years. Enriched uranium is extremely difficult to acquire; over the past decade, Mueller argues, there were only 10 known thefts. The material stolen weighed a combined 16 pounds, which was nowhere near the amount needed to build a bomb. Once the uranium is acquired, building the weapon is simple in theory (anti-nuclear activist Howard Morland published a famous 1979 article about this in the Progressive) but quite difficult in practice, which is why entire countries have had to work decades to acquire the bomb, only sometimes meeting with success. (Plutonium, another fissile material, is sufficiently dangerous and difficult to transport that nonproliferation experts seldom discuss it.)
Gathering material for a biological weapon may be somewhat easier, but actually fashioning that weapon would be harder, as witnessed by the fact that such weapons have scarcely ever been deployed, even by nations. On the rare occasions when they have been, they've failed to live up to their billing as weapons of mass destruction. "Perhaps the greatest disincentive to using biological weapons," John Parachini of the RAND Corporation testified before Congress in 2001, "is that terrorists can inflict (and have inflicted) many more fatalities and casualties with conventional explosives than with unconventional weapons." The same argument applies to chemical weapons. In theory, journalist Gregg Easterbrook has noted (citing a congressional report), under perfect conditions, one ton of sarin could kill up to 8,000 people. But it's "reasonably unlikely" that a terrorist group could acquire that much sarin, and perfect conditions mean no wind and no sun. Even light winds would reduce casualties to 800. You'd be better off detonating a conventional bomb in a city square.
On the other hand: Before 9/11 no self-respecting structural engineer would have predicted that you could reduce the World Trade Center towers and their inhabitants to dust by crashing two planes into them. The threat of an attack even more destructive than 9/11 is what risk analysts call "low-probability, high-risk." The likelihood is remote; the consequences would be devastating. Which probably makes the prospect as tempting to 9/11's murderous creators as finishing the Great American Second Novel was to Ralph Ellison.
Next: "The Flypaper Theory," in which we'll examine whether al-Qaida is (or was) too busy killing Americans in Iraq to kill them in the United States.