diff --git a/src/backends/neon/workloads/NeonConvolution2dWorkload.cpp b/src/backends/neon/workloads/NeonConvolution2dWorkload.cpp
index 6d30251..3f69718 100644
--- a/src/backends/neon/workloads/NeonConvolution2dWorkload.cpp
+++ b/src/backends/neon/workloads/NeonConvolution2dWorkload.cpp
@@ -43,6 +43,12 @@
     if (descriptor.m_BiasEnabled)
     {
         ARMNN_ASSERT(biases.has_value());
+        // Same for bias as weights. We don't currently support non const.
+        if (!biases.value().IsConstant())
+        {
+            return arm_compute::Status{arm_compute::ErrorCode::RUNTIME_ERROR,
+                                       "ArmNN NeonConvolution2dWorkload does not support non constant bias."};
+        }
         aclBiasesInfo = BuildArmComputeTensorInfo(biases.value(), descriptor.m_DataLayout);
         aclBiasesInfo.set_are_values_constant(biases.value().IsConstant());
         optionalAclBiasesInfo = &aclBiasesInfo;
@@ -91,6 +97,8 @@
         m_BiasTensor = std::make_unique<arm_compute::Tensor>();
         BuildArmComputeTensor(*m_BiasTensor, info.m_InputTensorInfos[2], m_Data.m_Parameters.m_DataLayout);
         m_BiasTensor->info()->set_are_values_constant(info.m_InputTensorInfos[2].IsConstant());
+        // We do not support dynamic bias
+        ARMNN_ASSERT(info.m_InputTensorInfos[2].IsConstant() == true);
     }
 
     arm_compute::PadStrideInfo padStrideInfo = BuildArmComputePadStrideInfo(m_Data.m_Parameters);
diff --git a/src/backends/neon/workloads/NeonDepthwiseConvolutionWorkload.cpp b/src/backends/neon/workloads/NeonDepthwiseConvolutionWorkload.cpp
index 2bccfa8..518faab 100644
--- a/src/backends/neon/workloads/NeonDepthwiseConvolutionWorkload.cpp
+++ b/src/backends/neon/workloads/NeonDepthwiseConvolutionWorkload.cpp
@@ -58,6 +58,12 @@
     if (descriptor.m_BiasEnabled)
     {
         ARMNN_ASSERT(biases.has_value());
+        // Same for bias as weights. We don't currently support non const.
+        if (!biases.value().IsConstant())
+        {
+            return arm_compute::Status{arm_compute::ErrorCode::RUNTIME_ERROR,
+                                       "ArmNN NeonDepthwiseConv2dWorkload does not support non constant bias."};
+        }
         aclBiasesInfo = BuildArmComputeTensorInfo(biases.value(), descriptor.m_DataLayout);
         aclBiasesInfo.set_are_values_constant(biases.value().IsConstant());
         optionalAclBiasesInfo = &aclBiasesInfo;
@@ -94,6 +100,8 @@
     {
         biasesPtr = &PolymorphicDowncast<IAclTensorHandle *>(m_Data.m_Inputs[2])->GetTensor();
         biasesPtr->info()->set_are_values_constant(info.m_InputTensorInfos[2].IsConstant());
+        // We do not support dynamic bias
+        ARMNN_ASSERT(info.m_InputTensorInfos[2].IsConstant() == true);
     }
 
     arm_compute::ITensorInfo* weightsInfo = weights.info();
diff --git a/src/backends/neon/workloads/NeonFullyConnectedWorkload.cpp b/src/backends/neon/workloads/NeonFullyConnectedWorkload.cpp
index 12eff36..ee3de80 100644
--- a/src/backends/neon/workloads/NeonFullyConnectedWorkload.cpp
+++ b/src/backends/neon/workloads/NeonFullyConnectedWorkload.cpp
@@ -38,6 +38,12 @@
     if (descriptor.m_BiasEnabled)
     {
         ARMNN_ASSERT(biases.has_value());
+        // Same for bias as weights. We don't currently support non const.
+        if (!biases.value().IsConstant())
+        {
+            return arm_compute::Status{arm_compute::ErrorCode::RUNTIME_ERROR,
+                                        "Arm NN NeonFullyConnectedWorkload does not support non constant bias."};
+        }
         aclBiases = BuildArmComputeTensorInfo(biases.value());
         aclBiases.set_are_values_constant(biases.value().IsConstant());
         optionalAclBiases = &aclBiases;
@@ -79,6 +85,9 @@
         m_BiasesTensorInfo = info.m_InputTensorInfos[2];
         BuildArmComputeTensor(*m_BiasesTensor, m_BiasesTensorInfo);
         m_BiasesTensor->info()->set_are_values_constant(m_BiasesTensorInfo.IsConstant());
+
+        // We do not support dynamic bias
+        ARMNN_ASSERT(m_BiasesTensorInfo.IsConstant() == true);
     }
 
     const arm_compute::ActivationLayerInfo activationInfo = ConvertAdditionalInfoToAclActivationLayerInfo(descriptor);