The possibility of AI consciousness can no longer be dismissed—and the ethical consequences may need to be considered before we know for certain. AI self-improvement is also gathering pace, as AI increasingly helps researchers build better AI, potentially accelerating its own development. In cybersecurity, hackers are becoming more powerful with AI even as its makers see defending against them as a potentially huge business. One argument for governing AI favors stronger security and clear liability over sweeping new regulation, holding those who deploy AI responsible for foreseeable harm. And China’s robot makers, having mastered increasingly sophisticated physical movement, are now racing to develop the AI “brains” that could make humanoid machines genuinely useful at work.