It's strangely summarized, but the full text (linked from https://www.researchgate.net/publication/344393003_A_neural_correlate_of_sensory_consciousness_in_a_c

It's strangely summarized, but the full text (linked from https://www.researchgate.net/publication/344393003_A_neural_correlate_of_sensory_consciousness_in_a_corvid_bird https://www.researchgate.net/publication/344393003_A_neural_... ) has more details. If there was a faint light, and the red target appears, the birds were taught (how?) to peck it. If there was light and a blue target appears, don't peck it. If there was no light, and the blue target appears, peck it, and if the target is red, don't peck it. So peck a red target = "I saw a light", peck a blue target = "I didn't see a light". If the opposite color target appears to what you need to peck, don't peck it. Or: Target is red, did I see a light? If yes, peck it. Target is blue. Was there a light? If there wasn't, peck it. I didn't continue reading the paper, but I guess the researchers concluded since they crows could answer the "Did I see a light?" question correctly most (i.e. statistically significant amount) of the times, it shows they can ponder the contents of their brains.