The bold experiment is proving attractive, but comes fraught with dangers.
In mid-December 2016, in rural Diffa region on Niger's southern border with Nigeria, fourteen men gave themselves up to authorities. The group said that they were former fighters of Boko Haram and that they had abandoned their weapons in the bush.
News of this impromptu surrender from the Islamist militant group responsible for tens of thousands of deaths and millions of displacements came as a surprise to most in the area. But not to regional authorities.
Since late last year, they had been quietly testing a tactic of asking families whose children have joined Boko Haram to spread word of an amnesty. If they surrendered, fighters were told, they would be pardoned and assisted in rejoining their communities.
Before then, the main regional response to the brutal Islamist militant group had been military. This has had some successes in weakening the combatants, and the last major Boko Haram attack in Niger in which civilians were killed was in September 2016. But in Nigeria, where the group originated, and beyond, gruesome assaults, abductions, and bombings of schools and markets continued.
To those in Diffa, these attacks have been shocking. But more distressing to many has been the rate of voluntary conscription amongst Niger's youth. Imams and village chiefs return to one question: "What about this savagery is attractive to our young?" Families and leaders tussle with this issue, but many simply refuse to countenance that those who join Boko Haram from Niger are truly radicalised.
It was with this belief in mind - as well as an awareness of the limits of a ground war - that the experimental amnesty plan was hatched last year. The exact details of the "secret messaging" campaign are unclear, but local leaders express pride in their initiative, which they say is ongoing, and follow it closely.
As the prefect of Maïné-Soroa told me, "Governor [of Diffa Region] Dan Dano calls every night to ask how many Boko have surrendered."
As of late-March, the number stood at nearly 150 across Diffa.
In terms of numbers, the amnesty scheme has so far proven to be effective. The logic behind it is also clear. Uganda's use of a similar strategy to entice defections from the Lord's Resistance Army in the early-2000s is widely believed to have weakened rebel ranks. And Diffa's experiment comes at a time when Boko Haram is already facing factional splintering and other difficulties.
As a locally-designed and -executed initiative, it is also impressive and promising. Often when disarmament, demobilisation and reintegration (DDR) schemes are implemented, they are imported internationally with little local ownership. But this is not the case with Diffa, and other regions facing the same problem are watching the bold experiment closely.
However, while local leaders are buoyed by progress so far, not everyone is convinced.
Some believe that the policy is a distraction from tackling the longer-term push factors - such as poverty and a weak state - that lead youth to join Boko Haram in the first place. Meanwhile others worry that funds from other more widely beneficial development projects will be re-directed to rehabilitating former combatants.
As Niger's Minister of Justice Marou Amadou says of ex-Boko Haram fighters, "it costs us money to house them, to feed them".
For his part, Governor Dano says he is not yet seeking funds to help manage the growing caseload. His intent is to pilot the idea and, if it proves tenable, to seek support where it is needed. But this more reactive approach also brings with it certain risks.
At present, anticipated needs only cover the Goudoumaria reintegration centre where vocational training and de-radicalisation programmes are to take place over a two-year period. As in combatant DDR programmes elsewhere, external partners will be involved.
However, if defector numbers spike with no clear plan or resources already in place, the programme could stall. Frustrations could escalate and deserters may revolt or even re-mobilise. This has happened in many other DDR programmes where logistics and planning were slow or inadequate.
Another serious challenge to the amnesty comes from the fact that, at a grassroots level, many local communities in Niger are not yet on board with the idea. They view the deserters with suspicion and hostility.
Unlike in Uganda, there is currently no legal framework for Diffa's amnesty initiative, meaning there is no official process by which ex-combatants can gain legal status as pardoned deserters. Moreover, some worry that those surrendering are being planted by Boko Haram.
Dano concedes that processing the defectors will take time, but insists there are measures in place to determine threat levels.
"We cross-reference their stories, their claim to a certain village and family, by visiting those places and confirming details. We try to learn more about them, when they left and if witnesses saw them attacking villages here," he says. He also suggests that those who are genuinely radicalised will simply ignore the offer of an amnesty.
In order to drive support for the initiative, Dano along with local prefects and leaders have been appearing before the public. But from all reports, these are purely declaratory rather than responsive exercises.
This could pose a serious problem. If local leadership fails to convince the population, it could undermine the whole endeavour. After all, it is ultimately victims - more so than ex-combatants or state officials - whose buy-in is essential for an amnesty to be effective. For reconciliation and reinsertion of former fighters to be possible, communities must be prepared to accept them back into their lives.
Yet there are currently no participatory approaches being adopted to more closely involve communities, and many simply see the amnesty as impunity. Furthermore, popular sentiment may harden as word spreads that deserters could be rewarded with vocational training and livelihoods assistance while innocent, traumatised communities get nothing.
"We think we are diminishing the ranks of BH with this amnesty effort, but now what are we doing with the defectors?" asks Minister Amadou. "We aren't prosecuting them - none of this is good for us."
The price of peace?
The challenges and risks of Diffa's pilot amnesty are thus clear to see. Trying to pardon and rehabilitate former fighters under volatile and uncertain circumstances comes fraught with dangers, especially if the initiatives are not carefully and thoroughly financed and planned.
Meanwhile, if local communities remain resistant to the idea, the policy could result in deepening resentment, hostility and suspicion.
However, as Boko Haram continues to terrorise Niger and the Lake Chad region, local authorities insist that the risks of continuing with a predominantly military approach are similarly grave.
"We cannot become Nigeria", says Dano.
Fighting Boko Haram may involve policies that are controversial to begin with, say local leaders, but they are ultimately necessary.
Asked how he justifies pardoning former Boko Haram militants and spending scarce funds on their rehabilitation to those in Diffa, the Maïné-Soroa prefect sighs. "I tell them such is the price we have to pay for peace".