Total yearly consumption of oil worldwide:
2.5e10 barrels oil = 3.97e15 grams,
gas: 9.5e13 cub feet = 6.65e31 molecules at atmospheric pressure = 3.97e15 grams of pure carbon
coal: 5.44e9 short tons = 4.94e15 grams
Total mass of carbon consumed 1.02e16 gram (assuming oil and coal are pure carbon)
Total mass of earth atmosphere: 5.1e21 gram
At 0.0368% mass concentration of CO2 gives roughly 6.26e17 grams of carbon (for the sake of quick calc mass of C,N and O atoms) as part of CO2 in atmosphere.
So if we assume that total worldwide yearly consumption of carbon in oil, coal and gas is entirely converted to CO2, then we have to say that humans produce early 1.64% of the total CO2 in the atmosphere. This is ball park figure, let us take +-2 half an order magnitude error:
0.5-5% of total CO2 in the atmosphere.
Now I can cautiously formulate what I think on the influence of humans on CO2.
Do we influence it remarkedly? This is scary part. In complete reversal of my earlier position, I say: yes.
Now we know that in the last 20 years annual CO2 increase was 1.5ppm per year which is 0.3% of the total CO2 in the atmosphere.
The changes in the increase of CO2 in the atmosphere are comparable to the emissions! One can do even wilder assumtion that "nature" is "trying" to reduce CO2 by 1.3% a year, while "humans" are "overcoming" this resulting in increase of CO2 by 0.3% a year, but I am not going to do that.
Now the second question:
Can we do something significant about it? No.
Kyoto protocol proposes to cut emissions by 5.3%. Even if we assume that 0.5-5% of total CO2 make the difference, 0.025-0.25% of total CO2 in the atmosphere hardly will make a difference.
After doing my assessment I found on the web a different LOWER figure for carbon emissions: 6.50E+15 grams of carbon which is 30% than my assessment. 30% is not a bad error for a small lab work.