Dyes have electron delocalization. That's why they're optically active. Those bonds will gladly participate in chemical reactions within your body. (One of the cited chemicals was a biphenyl, and looked particularly nasty.)
Chemical straightener is even worse. They're intended to break disulfide bonds, which are of critical importance in biochemical structure.
This stuff could percolate to your DNA and introduce deleterious changes.
What do you mean by that?
When electrons are tightly bound into nodes in and between the vertices, they are energetically ‘stable’. They, therefore, do not like change.
When electrons are shared equally between all nodes and are mobile, they could be considered to be ‘more active’. As the electrons are not tightly bound, it is ‘all to play for’ and some change in state may be accommodated.
The optical activity described occurs because the equally shared, ‘central node’ electrons are ‘free-er’ and have the flexibility to accept unusual states. This is because in the multi-node model, the electrons may be considered tightly bound to each node and so any the effect of any change in the state of a given electron is highly concentrated and local.
In the shared electron central node model the electrons are ‘more mobile’ and so any change in state has a distributed and shared effect across all electrons, rather than only those in a given node of the multi modal model.
To emit light, an electron must absorb enough energy to reach ‘the next level’. In the multi nodal model, the next level is extremely far away and may require so much energy that the molecule will disintegrate rather than emit light.
In the equally shared central node model, the next level is ‘accessible’ in that, whilst unstable, the mobile bulk of electrons in the molecule can accommodate the change and radiate the excess away.
The exact colour of light is determined by the energy required to reach the next level, in most cases. This is a consequence of the delocalisation of electrons in dyes, which may be considered as following the shared central node model. A further consequence is increased reactivity.
This explanation should be considered poor and ignores a lot of nuance and theory (whilst relying on a hazy memory). Likewise, it neglects to explain intriguing phenomena such as the anion-pi interaction. In any case, I hope it is not ‘too wrong’ and helps!