The flaw in your logic comes from neglecting the water's acceleration after cutting the string, and forces on the water before cutting the string. The ball is accelerating upward, yes, and that would seem to indicate that more force needs to be supplied from the scale. But the center of mass of the system (ball, water, and container) is accelerating down. Of course it is, that's the whole point of the buoyant force - there is less potential energy when the ball is replaced with water and the ball moves up because the water is heavier. Like I said in a comment, this is analogous to what is shown in this video. This is a bit paradoxical - it visually looks like something is accelerating up, but rest assured more mass is accelerating down than is accelerating up. The buoyant force is not in the business of increasing the gravitational potential of the whole system. We expect the reading on the scale to be lower in the moments after the string is cut, and that's exactly what we'll find after calculating the size of the effect properly.
Before the string is cut the weight measured on the scale is definitely the entire mass of the system times $g$, as it is for all static systems (that's what scales are for). One might think the weight on the scale is reduced by the tension in the string but it definitely is not. To see why, consider this more careful accounting of the forces on the container, the ball, and the water. Ultimately, the buoyant force that would seem to pull up on the container is cancelled by extra force in the water that is also transferred to the container.
![enter image description here](https://cdn.statically.io/img/i.sstatic.net/APw8p.png)
Likewise if a person stands on a scale and pulls up really hard on a handle connected to the scale, the scale still measures exactly the person's mass times $g$. The force that the person pulls up is countered by additional force in his feet, which is also directed into the scale.
![enter image description here](https://cdn.statically.io/img/i.sstatic.net/67fMC.png)
So now what about the situation after cutting the string? Like I said we need to think about the water accelerating down, not just the ball moving up. To make the geometry more trivial my ball is a square, but rest assured the answer in the end does not depend on the shape of the ball.
![enter image description here](https://cdn.statically.io/img/i.sstatic.net/zPmLA.png)
In a short time $t$ the "ball" (side length $h$) moves up by a distance $(1/2)at^2$. As you correctly calculated $a=(F_B-W)/m$. Water is transferred from above the ball to below, through height $h$. A total volume $(1/2)ah^2t^2$, or a mass of $(1/2)at^2h^2\rho_W$ ($\rho_W$ is the density of water), moves down by $h$. Equivalently, a mass of water $\rho_W h^3$ accelerates downward at a rate $a$.
So finally the answer... the container pushes up on the water enough that the rest of the water not accelerating downward remains still. The forces on the water are $gM_w$ pulling down, and $B$ (the buoyancy of the ball) pushing down (the reaction force of the buoyant force on the ball). So the total force that the container exerts on the water, and thus the reading on the scale, is
$$
gM_w+B-\rho_W h^3 a
$$
But this can be made nicer by inserting, for example, $B=\rho_W Vg$ ($V$ is now the generic volume of a ball) and $a=(B-M_b g)/M_b$, $M_b=\rho_B V$ (mass of the ball is the density of the ball times volume)
$$
gM_w+\rho_W V g-\rho_W V(\rho_W Vg-\rho_B V g)/\rho_B V
$$
$$
gM_w+Vg(2\rho_W- \rho_W^2/\rho_B)<gM_w+\rho_B Vg
$$
This is always lower than the initial reading $gM_w+\rho_B Vg$. $2-1/x<x$ except when $x=1$. I.e. obviously except when the ball is the same density as water there is no change. This answer even shows that when the ball is heavier (and presumably pushed up by a stiff rod that is removed), the weight goes down as well, but that's more obvious.
Interesting to see that after the string is cut, whether the weight is higher or lower than the mass of the water alone depends on if the density of the ball is less than half of the density of water.
Another amazing feature of this solution is that the mass drop goes to infinity as $\rho_B$ goes to zero. This is somewhat real - when the mass of the ball goes to zero it accelerates up infinitely fast. A lot of water is transported down really quickly to compensate. You might think it's crazy for water to accelerate down faster than $g$, but the acceleration of a small volume of water near the ball is driven by the pressure of all the water above it more than it's driven by the effect of gravity on that specific fluid volume. Probably this accelearation is very short-lived; it's limited by viscosity. Of course we're more familiar with things floating up at roughly constant speed, because they reach terminal velocity really fast.