they will have to pay the workers more to entice them in, and provide better working conditions.
this will generate more work, and higher consumption.
The problem is, noone has tried to see the complete picture by adding up these effects. The net effect is probably negative, unless this is introduced all over the world simultaneously.
Businesses will hire more people and pay them more => prices will go up, purchasing power for domestic products will go down (i.e. the same income will afford fewer domestic products). Imports will go up because those products will be cheaper than before in comparison. Exports will go down for the same reason.
Therefore, a large part of the expected higher consumption will go to imports / foreign profits. Due to reduced exports, domestic production will go down instead of up. Since jobs cost more now, there will be even fewer of them than expected.
Overall, countries which introduce this, will suffer.
It would probably be worthwhile to give people part money and part vouchers for certain eligible (i.e. domestic but some international treaties might not allow this discrimination of imported goods) products, in order to keep the purchases of domestic goods at a high level.
Your Keynesian story may eventually play out - the real value of the BI might be inflated away, resulting in people going back to work. Perhaps that will completely mitigate the problem, or perhaps it will merely result in a new equilibrium which only hax 0 < X% < 9% fewer people working. But why create the problem in the first place?
If 100 people produce 10 apples each, you have 1000 apples. If 9 of them quit, you have only 910 apples.
If 100 people produce 10 apples each, and 99 robots take over the job of producing apples, there will be 99 unemployed people with no money to buy your 1000 apples.
I don't think we have good information on second and third order effects.
Personally, I think it isn't an impossible problem to find a way to provide a system where people get food and shelter, but are still given incentives to work. Most people do a lot of things for reasons that have nothing to do with food and shelter. With the right cultural values in place, people will still work. (For example, we could implement BI using dogecoin, so people are shamed into going to work.)
Eventually after carving away vast segments of the population, you're left with unabomber style nuts, hippies, the economically useless, and the disabled, none of whom work anyway. It would be decentralized and self selected rather than centrally controlled and fairly inefficient, so that saves quite a bit of money and human suffering right there.
If there's fewer and fewer jobs every year, and that seems to be the case for quite awhile, then the problem with people not needing to work jobs that don't exist, would be...
Source? Widespread BI is usually proposed alongside abolishment of the minimum wage. This means that you'd expect wages at the low end to start reflecting job comfort more - crappy jobs (trash pickup?) would have to be paid more to entice people, but cushier jobs may well see pay go down.
That money didn't come from nowhere. There is an equal decrease in purchasing power for others due to the increased taxes needed to fuel such an expensive scheme.
the 9% 'drop in work' will be offset by the increase in consumption caused by a large range of people suddenly having more purchasing power,
businesses will have to increase production, and hire more workers to do so.
they will have to pay the workers more to entice them in, and provide better working conditions.
this will generate more work, and higher consumption.